# Numonic AI > Numonic is an AI-native digital asset management platform. It captures the full generation envelope around every AI asset — prompt, model, seed, workflow graph, lineage, and compliance metadata — so creative teams can find, reproduce, and prove the provenance of work made with ComfyUI, Midjourney, Stable Diffusion, and other generative tools. ## Product & company - [Numonic — AI-native digital asset management](https://www.numonic.ai/): Home: the AI creative system of record — capture, organise, and prove the provenance of generative AI assets. - [Pricing](https://www.numonic.ai/pricing): Free for creators. Teams pay for seats, storage, and AI credits. Every plan, Free included, gets the full product: library, lineage, prompt versions, REST API and MCP server. - [Product overview](https://www.numonic.ai/product): How Numonic captures the generation envelope: metadata, lineage, semantic search, collections, collaboration, and privacy-safe export. - [ComfyUI workflow management](https://www.numonic.ai/product/comfyui): Version ComfyUI workflows, preserve PNG metadata, and track team collaboration with compliance-ready provenance. - [Midjourney library management](https://www.numonic.ai/product/midjourney): Organise Midjourney outputs with prompt, parameter, and evolution-chain tracking across thousands of generations. - [Asset lineage](https://www.numonic.ai/product/lineage): Navigate the full chain of custody for every asset — from prompt through every variation, upscale, and edit. - [Semantic search](https://www.numonic.ai/product/search): Find assets by meaning, visual similarity, prompt, model, or metadata across the whole library. - [Solutions for solo creators](https://www.numonic.ai/solutions/creators): Tame thousands of AI outputs as a single operator, with the same API and MCP access as teams. - [Solutions for studios](https://www.numonic.ai/solutions/studios): Multi-client asset organisation, lineage tracking, and audit trails for creative studios. - [Six-Week Studio Pilot](https://www.numonic.ai/solutions/pilot): A paid, hands-on six-week pilot for creative studios of 2 to 15 people using generative AI. Measures hours per approved piece on real client work, with diagnostic measures and a written report. - [Solutions for enterprise](https://www.numonic.ai/solutions/enterprise): One library across brand, in-house studio, and agency teams, with an append-only audit log and full REST API and MCP access. SSO and data residency are on the roadmap. - [Solutions for compliance](https://www.numonic.ai/solutions/compliance): EU AI Act Article 50 and California SB 942: IPTC 2025.1 AI-disclosure fields written on export, C2PA credential detection on ingest, C2PA signing of JPEG, PNG and WebP exports, and an append-only log of AI content events. - [What Numonic marks and signs on export](https://www.numonic.ai/c2pa-export-formats): Format-by-format table of which exports get IPTC 2025.1 AI-disclosure metadata and a signed C2PA manifest today (PNG, JPEG, WebP; GLB for disclosure fields), which do not yet (video, audio, GIF), and what is planned. - [Company](https://www.numonic.ai/company): About Numonic Labs and the team building the AI creative system of record. - [Trust center](https://www.numonic.ai/trust): Security, privacy, and data-handling commitments across all plans. - [Blog](https://www.numonic.ai/blog): Articles on AI asset management, creative workflows, provenance, and compliance. - [Glossary](https://www.numonic.ai/glossary): Definitions for AI asset management, content provenance, and regulatory compliance terms. - [Guides](https://www.numonic.ai/guides): In-depth pillar guides to organising and governing AI-generated content. - [Creative production metrics for AI work](https://www.numonic.ai/resources/creative-production-metrics): The one number each type of creative team should track for AI work (e.g. hours per approved piece), how the numbers connect, and how to measure reuse rate. - [Free tools](https://www.numonic.ai/tools): Free interactive utilities: metadata inspectors, an EU AI Act disclosure generator, and library calculators. - [Integrations](https://www.numonic.ai/integrations): ComfyUI node reference and integration guides for the generative tools Numonic supports. - [API reference](https://www.numonic.ai/docs/api): REST API reference for Numonic — authentication, endpoints, and the OpenAPI spec at /api/v1/openapi.json. - [MCP server reference](https://www.numonic.ai/docs/api/mcp): Connect an AI agent to Numonic over MCP — tool reference and the machine-readable spec at /api/v1/mcp-spec.json. - [Connect-an-agent quickstart](https://www.numonic.ai/docs/api/mcp/quickstart): Connect an MCP-capable agent with an API key or through OAuth; server card at /.well-known/mcp/server-card.json. ## Guides (full summaries) ### The Complete Guide to ComfyUI Asset Management Master ComfyUI asset management: workflow versioning, PNG metadata preservation, team collaboration, and compliance-ready provenance tracking. https://www.numonic.ai/guides/comfyui-asset-management ### How to Organise AI-Generated Images: The Complete Guide The solo creator’s guide to taming thousands of AI outputs. Learn naming conventions, folder systems, tagging strategies, and multi-tool workflows. https://www.numonic.ai/guides/organise-ai-generated-images ### AI Content Compliance for Agencies: From Governance Bottleneck to Competitive Advantage The agency guide to EU AI Act Article 50, California SB 942, IPTC 2025.1, C2PA metadata, and compliance workflows that accelerate creative velocity. https://www.numonic.ai/guides/ai-content-compliance ### AI-Native DAM Architecture: A Pattern Language for Generative Asset Management A guide to building DAM systems for AI-generated content — search, lineage, compliance, and curation patterns. https://www.numonic.ai/guides/ai-native-dam-architecture ### Organise Your Midjourney Library: The Complete Post-Export Guide Master Midjourney image organisation beyond the web app. Export strategies, metadata preservation, legacy back-catalogue recovery, style reference management, and compliance-ready workflows for 5,000+ image libraries. https://www.numonic.ai/guides/midjourney-organisation ### C2PA and Content Credentials for AI Creative Work What C2PA and Content Credentials record, why they vanish from AI images, and how Numonic signs your exports so the record travels with the file. https://www.numonic.ai/guides/c2pa-content-credentials ### ComfyUI Workflows: Save, Find and Recover Them Where ComfyUI stores the workflow behind every output, how to load it back, and how to search thousands of outputs by model, seed or sampler. https://www.numonic.ai/guides/comfyui-workflows ## Glossary (full definitions) ### AI Content Provenance A verifiable record of an AI-generated asset's origin, including the model, prompt, parameters, and workflow used to create it. Provenance enables reproducibility, compliance with regulations like the EU AI Act, and proof of creative ownership. AI content provenance goes beyond simple metadata. It encompasses the entire chain of custody for a generated asset—from the initial prompt and model configuration through every variation, upscale, and edit. For creative teams, provenance answers three critical questions: Who created this? How was it created? Can it be reproduced? Without provenance infrastructure, organizations face mounting risks: inability to prove asset origins during audits, duplicate work when parameters are lost, and compliance gaps as regulations like the EU AI Act Article 12 mandate detailed logging of AI system operations. https://www.numonic.ai/glossary/ai-content-provenance ### C2PA The Coalition for Content Provenance and Authenticity (C2PA) is an open technical standard for certifying the origin and history of digital content. It embeds cryptographic signatures into files to verify whether content is human-created, AI-generated, or edited. C2PA was founded in 2021 by Adobe, Arm, the BBC, Intel, Microsoft and Truepic to address misinformation and content authenticity at scale. The standard uses cryptographic manifests embedded directly in media files (images, video, audio and documents), creating a tamper-evident record of how content was created and modified. For AI-generated content, C2PA manifests typically record the generating tool and that the content is AI-generated (the digital source type). Prompts and generation parameters usually live in the tool’s own metadata, not the manifest. This makes C2PA the emerging technical foundation for AI transparency regulations. Major platforms including Google, Meta, and OpenAI have committed to supporting C2PA verification. https://www.numonic.ai/glossary/c2pa ### AI Metadata Structured information embedded in or associated with AI-generated content that describes how the content was produced. This includes model identifiers, prompt text, generation parameters (seed, steps, CFG scale), and workflow configurations. AI metadata is the technical foundation for provenance, compliance, and reproducibility. Different AI tools store metadata differently: ComfyUI embeds workflow JSON in PNG tEXt chunks, Midjourney stores parameters in Discord message text, and Stable Diffusion UIs use various EXIF and PNG metadata fields. The challenge is that AI metadata is fragile. Social media platforms strip it on upload, image editors discard it during saves, and many workflow tools never capture it in the first place. An AI-native DAM system must extract, normalize, and preserve this metadata across the asset lifecycle. https://www.numonic.ai/glossary/ai-metadata ### Workflow Reproducibility The ability to recreate an AI-generated asset by replaying its original workflow with identical parameters. Reproducibility requires preserving the complete generation context: model version, seed value, sampler settings, and node configurations. Reproducibility is a cornerstone of professional AI creative work. When a client requests a variation of a generated image, the creator must be able to reproduce the original result exactly before making modifications. Without reproducibility, creative teams resort to trial and error—wasting time and compute resources. In ComfyUI, reproducibility depends on preserving the full workflow JSON (node graph, connections, and per-node parameters) alongside the generation seed. Even minor differences in model versions or sampler implementations can produce different outputs from identical parameters. https://www.numonic.ai/glossary/workflow-reproducibility ### DAM (Digital Asset Management) A system for organizing, storing, retrieving, and distributing digital files. Traditional DAM platforms manage photos, videos, and documents with metadata tagging, version control, and access permissions. AI-native DAMs extend this with provenance tracking and workflow capture. Digital Asset Management has existed since the early 2000s, primarily serving marketing teams, media companies, and enterprises managing large libraries of brand assets. Traditional DAMs like Bynder, Brandfolder, and Adobe Experience Manager focus on human-created content with manual tagging workflows. The explosion of AI-generated content has exposed fundamental limitations in traditional DAM architecture. These systems were not designed to handle the volume (thousands of assets per day), the metadata richness (workflow JSON, generation parameters), or the provenance requirements (audit trails, regulatory compliance) that AI content demands. https://www.numonic.ai/glossary/dam ### EU AI Act The European Union's comprehensive regulation governing artificial intelligence systems. Article 12 requires providers of high-risk AI systems to maintain detailed logging and audit trails. The Act establishes transparency requirements for AI-generated content, including disclosure obligations. The EU AI Act, which entered into force in August 2024 with phased compliance deadlines through 2027, creates a risk-based regulatory framework for AI systems. For creative professionals and marketing agencies, the most relevant provisions include: Article 12 mandates automatic logging of AI system operations, creating audit trail requirements that affect anyone deploying AI in production. Article 50 requires clear disclosure when content is AI-generated. The penalties for non-compliance reach up to 35 million euros or 7% of global annual turnover. Organizations using AI for content creation need infrastructure to capture, store, and retrieve provenance data to demonstrate compliance during audits. https://www.numonic.ai/glossary/eu-ai-act ### ComfyUI Workflow A node-based visual programming graph in ComfyUI that defines the complete image generation pipeline. Each node represents an operation (model loading, sampling, conditioning, post-processing), and connections between nodes define the data flow from input to output. ComfyUI workflows are stored as JSON objects that describe every node, connection, and parameter in the generation graph. When a PNG image is generated, ComfyUI embeds two key metadata chunks: the "workflow" (full node graph for the UI) and the "prompt" (flattened execution parameters). These workflow JSON objects are the most complete provenance records available in the AI image generation ecosystem. They capture not just the prompt text and seed, but the exact model, VAE, LoRA weights, sampler, scheduler, and every post-processing step. Preserving workflow JSON is essential for reproducibility and compliance. https://www.numonic.ai/glossary/comfyui-workflow ### AI-Native DAM A digital asset management system designed from the ground up for AI-generated content. Unlike traditional DAMs adapted for AI workflows, an AI-native DAM automatically extracts metadata, tracks provenance, enables workflow reproducibility, and supports compliance requirements. The distinction between "AI-enabled" and "AI-native" is architectural. Traditional DAMs adding AI features (auto-tagging, smart search) are AI-enabled—they use AI to enhance existing workflows. An AI-native DAM is built with AI content as the primary use case, meaning provenance tracking, workflow capture, and parameter preservation are core features rather than add-ons. Key capabilities of an AI-native DAM include: automatic metadata extraction from AI tool outputs, semantic search across generation parameters, lineage tracking showing how assets evolved through variations and edits, and compliance-ready export with embedded provenance data. https://www.numonic.ai/glossary/ai-native-dam ### Prompt Metadata The text prompts, negative prompts, and associated generation settings captured alongside an AI-generated asset. Prompt metadata enables search, categorization, and reproduction of creative outputs across teams and workflows. Prompt metadata encompasses more than the user-visible text prompt. It includes negative prompts (what to avoid), prompt weighting and emphasis syntax, embedding references (textual inversions, LoRA triggers), and the prompt encoding parameters (CLIP skip level, token limit handling). In practice, prompt metadata is scattered across different locations depending on the tool: ComfyUI stores it in PNG tEXt chunks, Automatic1111 uses PNG info text and EXIF, and Midjourney embeds it in the PNGs. Normalizing and preserving prompt metadata across these sources is a core challenge for AI asset management. https://www.numonic.ai/glossary/prompt-metadata ### AI Audit Trail A chronological record of all operations performed by or with an AI system, including inputs, outputs, configuration changes, and user interactions. Required by EU AI Act Article 12 for high-risk AI systems and increasingly expected for all commercial AI use. An AI audit trail serves both compliance and operational purposes. For compliance, it provides the documentation regulators require to verify that AI systems are operating within legal boundaries. For operations, it enables debugging, quality assurance, and accountability when AI outputs are questioned. Effective audit trails for AI-generated content must capture: who initiated the generation, what model and parameters were used, when the generation occurred, what the output was, and how the output was subsequently used or modified. The challenge is building this capture infrastructure without creating friction that slows creative workflows. https://www.numonic.ai/glossary/ai-audit-trail ### IPTC 2025.1 The 2025.1 revision of the IPTC Photo Metadata Standard that introduces dedicated fields for AI-generated content. New fields include AISystemUsed, AISystemVersionUsed, and DigitalSourceType, providing a standardized vocabulary for recording how AI tools contributed to content creation. The International Press Telecommunications Council (IPTC) has maintained photo metadata standards since 1979. The 2025.1 revision is significant because it formalizes AI-specific fields that integrate with existing XMP and EXIF metadata infrastructure already supported by major software platforms. For creative agencies, IPTC 2025.1 matters because it provides a vendor-neutral way to record AI involvement that survives across tools and platforms. Unlike proprietary metadata formats, IPTC fields are recognized by Adobe Creative Suite, DAM systems, and stock photography platforms. The DigitalSourceType field distinguishes between human-created, AI-generated, and AI-assisted content—a distinction increasingly required by regulators and clients. https://www.numonic.ai/glossary/iptc-2025-1 ### Shadow AI The use of AI tools by employees without organizational knowledge, approval, or governance. Similar to shadow IT, shadow AI creates compliance risks when staff use unapproved generative tools for client work without proper metadata tracking, disclosure, or IP verification. Shadow AI is particularly prevalent in creative agencies where individual designers and copywriters adopt AI tools for productivity without formal approval. A 2026 industry survey found that over 60 percent of creative professionals use AI tools not sanctioned by their employer for at least some client work. The risks of shadow AI extend beyond compliance. Without governance, agencies cannot guarantee that AI-generated deliverables carry proper provenance metadata, meet client disclosure requirements, or avoid training-data IP issues. Shadow AI also prevents accurate billing—agencies may undercharge for AI-assisted work or overcharge for fully AI-generated outputs without realizing it. Addressing shadow AI requires a combination of clear policy, approved tool registries, and technical infrastructure that makes compliant AI usage easier than unsanctioned alternatives. https://www.numonic.ai/glossary/shadow-ai ### AI Deployer Under the EU AI Act, a natural or legal person that uses an AI system under its authority, as distinct from the AI provider who develops or markets the system. Most creative agencies are deployers, which carries specific transparency, monitoring, and record-keeping obligations. The EU AI Act distinguishes between AI providers (companies like OpenAI, Stability AI, and Midjourney that build the models) and AI deployers (organizations that use those models in their operations). This distinction matters because each role carries different obligations under the regulation. For creative agencies, deployer obligations include: ensuring AI systems are used in accordance with their instructions, implementing human oversight measures, monitoring AI system operations for risks, and maintaining logs of AI system use. Agencies that fine-tune models or build custom workflows may cross into provider territory, triggering additional requirements. Understanding your classification—deployer, provider, or both—is the first step in building an effective compliance programme. https://www.numonic.ai/glossary/ai-deployer ### SB 942 (California AI Transparency Act) California Senate Bill 942, effective January 2026, requiring providers of generative AI systems to offer provenance tools and users of covered AI systems to disclose AI-generated content in commercial contexts. Applies to systems with over one million monthly users. SB 942 complements the EU AI Act by establishing US-based transparency requirements for generative AI. The law requires AI providers to build provenance detection tools and make them freely available. For creative agencies operating in or serving California clients, SB 942 creates disclosure obligations when AI-generated content is used commercially. The practical impact for agencies includes: labeling AI-generated content in advertising and marketing materials, maintaining records of which deliverables involved AI tools, and ensuring clients are informed about AI usage in their projects. AB 853, a companion bill, extends similar requirements to political advertising. Together, these laws signal that California is following the EU in mandating AI content transparency. https://www.numonic.ai/glossary/sb-942 ### AI Content Disclosure The practice of transparently communicating to audiences that content was created with or by artificial intelligence. Disclosure requirements vary by jurisdiction—mandated by EU AI Act Article 50, California SB 942, and increasingly expected by platform policies and industry standards. AI content disclosure sits at the intersection of legal compliance, ethical practice, and client trust. The EU AI Act Article 50 requires that persons exposed to AI-generated content are informed of its artificial origin, with limited exceptions for artistic and satirical use. California SB 942 adds US-based requirements for commercial contexts. For creative agencies, disclosure requires both policy and infrastructure. Policy determines when and how to disclose—in contracts, on deliverables, in published media. Infrastructure ensures the technical capability exists to track which assets are AI-generated, embed provenance metadata, and produce audit documentation on demand. Agencies that build disclosure into their standard workflow gain a competitive advantage as clients increasingly request transparency guarantees. https://www.numonic.ai/glossary/ai-content-disclosure ### Content Credentials A consumer-facing implementation of the C2PA standard that displays a tamper-evident provenance record for digital content. Content Credentials show whether an image, video, or audio file was created by a human, generated by AI, or edited, along with the tools used. Content Credentials are the user-visible layer of the C2PA technical standard. While C2PA defines the cryptographic manifest format, Content Credentials provide the interface through which viewers can inspect an asset's history. Adobe's Content Authenticity Initiative popularized the concept, and major platforms including Google, Meta, and Microsoft now support Content Credentials verification. For creative agencies, Content Credentials serve a dual purpose. They satisfy regulatory disclosure requirements by embedding machine-readable provenance directly in deliverables. They also build client trust by providing verifiable proof of creative process. As AI-generated content becomes harder to distinguish visually from human-created work, Content Credentials become the primary mechanism for establishing authenticity and creative attribution. https://www.numonic.ai/glossary/content-credentials ### Content-Addressed Storage A storage model that identifies files by a cryptographic hash of their content rather than by file path or name. Two files with identical binary content produce the same hash and are stored once, eliminating duplicates regardless of how many times they are imported or where they appear in a library. Content-addressed storage inverts the traditional relationship between identity and location. In path-based storage, a file's identity is its name and folder — rename it and you lose the reference. In content-addressed storage, the file's identity is its content hash — move it, rename it, or import it again, and the system recognizes it as the same asset. For AI-generated asset libraries, where the same image may be saved from multiple tools under different names or imported alongside workflow outputs multiple times, content-addressed storage eliminates the 20-30% redundancy that accumulates in path-based systems. It also provides the foundation for collection branching, where lightweight copies can reference the same underlying files without duplication. https://www.numonic.ai/glossary/content-addressed-storage ### Ingest Pipeline The multi-stage processing system that transforms a raw uploaded file into a fully indexed, searchable asset. Stages include content hashing for deduplication, tool-specific metadata extraction, thumbnail generation, and asynchronous enrichment such as embedding generation and session clustering. The ingest pipeline is the central architectural component of an AI-native DAM. Every file that enters the system — whether uploaded manually, synced from ComfyUI, or received through an API — passes through the same pipeline stages. This guarantees consistent metadata extraction and indexing regardless of how the file arrived. The pipeline is designed as a series of stages with increasing cost and latency. Cheap operations (hashing, deduplication, thumbnail generation) run synchronously during upload. Expensive operations (embedding generation, quality scoring, session clustering) run asynchronously in the background, progressively enriching the asset record over minutes rather than blocking the upload for seconds. https://www.numonic.ai/glossary/ingest-pipeline ### Embedding Space A high-dimensional mathematical space where images and text are represented as numerical vectors such that semantically similar content occupies nearby positions. The foundation of semantic search — querying by meaning rather than exact keywords — with cosine similarity measuring relatedness between vectors. Embedding spaces are created by neural networks trained to map visual and textual content into fixed-length vectors (typically 512 or 1536 dimensions). The key property is that similar concepts cluster together: a vector for "cyberpunk street scene" will be close to "neon rain-slicked alley, Blade Runner aesthetic" despite sharing zero keywords. For AI asset management, embedding spaces enable two critical capabilities. First, semantic search lets users find images by describing what they look like rather than remembering exact tags or filenames. Second, near-duplicate detection identifies perceptually identical images that differ at the binary level — different seeds, minor edits, or format conversions that content hashing would miss. https://www.numonic.ai/glossary/embedding-space ### Hybrid Search A retrieval architecture that combines structured metadata queries (exact filters on tool, date, model) with vector similarity search (semantic meaning, visual resemblance) in parallel, then fuses the ranked results into a single list. Enables queries that are simultaneously precise and semantic. Neither keyword search nor semantic search alone is sufficient for AI asset libraries. Keyword search excels at exact matches ("SDXL, seed 42") but fails when users describe visual concepts. Semantic search understands meaning but cannot enforce strict constraints like date ranges or tool names. Hybrid search runs both retrieval paths in parallel and combines their ranked results using fusion algorithms. A query like "moody cyberpunk portrait, made with SDXL last week" splits into a structured filter (tool=SDXL, date=last 7 days) and a semantic query ("moody cyberpunk portrait"), with results merged to satisfy both constraints simultaneously. https://www.numonic.ai/glossary/hybrid-search ### Creative Session Clustering The automatic grouping of AI-generated assets into discrete creative sessions by detecting boundaries from temporal gaps, parameter changes, and tool switches. Provides retroactive organizational structure without manual tagging — a session boundary is inferred when the generation rhythm pauses. Creative professionals rarely generate images one at a time. They work in focused sessions — exploring a concept, iterating on variations, adjusting parameters — then pause before starting something new. Session clustering detects these natural rhythms by analyzing timestamps, parameter continuity, and tool context. The organizational value is significant: sessions become a middle layer between individual assets and the entire library. Users can browse by session ("Tuesday afternoon's character explorations") rather than scrolling through a flat chronological list. Sessions also provide context for lineage tracking — variations within a session are likely related even if their generation parameters differ substantially. https://www.numonic.ai/glossary/creative-session-clustering ### Metadata Inversion The architectural shift where AI-generated assets arrive with rich generation metadata already embedded by the creation tool, inverting the traditional DAM assumption that metadata must be added by humans after upload. The challenge shifts from metadata creation to metadata capture and normalization. Traditional DAM systems are built on the assumption that files arrive with minimal metadata and humans must add descriptive information through manual tagging workflows. AI generation tools invert this: a ComfyUI output contains a complete workflow graph with every parameter, model reference, and processing step encoded in the PNG metadata. This inversion has profound architectural consequences. The bottleneck is no longer getting metadata into the system — it is extracting, normalizing, and making searchable the metadata that already exists. A system designed for metadata-scarce content (manual tagging) is fundamentally different from one designed for metadata-rich content (automated extraction and normalization). https://www.numonic.ai/glossary/metadata-inversion ### Metadata Normalization Pipeline A three-stage processing layer that translates tool-specific metadata formats from different AI tools (ComfyUI JSON, Midjourney Discord strings, DALL-E API responses) into a unified canonical schema. Enables cross-tool search and comparison from a single query interface. The normalization pipeline operates in three stages. Stage 1 (extraction) uses tool-specific parsers to read each format — ComfyUI PNG text chunks, Midjourney Discord messages, Stable Diffusion plaintext parameters. Stage 2 (mapping) translates tool-specific fields into canonical equivalents — "prompt" means different things structurally in each tool but refers to the same concept. Stage 3 (enrichment) adds derived information: content hashes, embeddings, session boundaries. The key architectural benefit is that downstream systems (search, lineage tracking, compliance export) only need to understand the canonical schema. Adding support for a new AI tool requires only a new Stage 1 extractor, not changes to every downstream consumer. https://www.numonic.ai/glossary/metadata-normalization-pipeline ### The Two Metadata Problem The absence of a common metadata standard across AI generation tools. Each tool uses its own storage location, format, field names, and encoding — ComfyUI stores JSON in PNG chunks, Midjourney uses Discord message strings, DALL-E returns API response objects — making cross-tool search architecturally difficult. The name refers to the fact that even a simple concept like "which model generated this image" has completely different representations across tools. ComfyUI records a file path to a checkpoint in a JSON workflow graph. Midjourney records a version flag (--v 6.1) in a text string. DALL-E records a model name in an API response. Same semantic meaning, entirely different structural representation. The two metadata problem is not just about formats — it extends to metadata richness. ComfyUI provides over 100 fields per generation. Midjourney provides 5-10. DALL-E provides 3-5. A normalization pipeline must accommodate this asymmetry without losing the rich information from detailed sources or failing on sparse sources. https://www.numonic.ai/glossary/two-metadata-problem ### Generation Lineage The chain of parent-child relationships between AI-generated assets — from an initial generation through variations, upscales, and refinements. Unlike code version control with explicit commits, image lineage must be inferred from parameter mutations, temporal proximity, and visual similarity. Generation lineage is harder to track than code version history because AI image creation lacks explicit save points. A developer commits code deliberately; an AI artist generates dozens of variations in rapid succession with no inherent relationship markers between them. The system must infer that image B is a variation of image A by comparing their generation parameters, creation timestamps, and visual similarity. Cross-tool lineage adds another dimension of difficulty. When an image generated in ComfyUI is edited in Photoshop and then fed back into Midjourney for style transfer, the provenance chain crosses three tools — each with different metadata formats and no awareness of what preceded it. Maintaining a continuous lineage record across these transitions is an open architectural challenge. https://www.numonic.ai/glossary/generation-lineage ### Collection Branching Creating lightweight copies of asset collections that maintain lineage to the original without duplicating underlying files. Analogous to Git branches for creative projects — enables parallel workstreams, versioned client deliveries, and safe experimentation while avoiding file multiplication. Collection branching leverages content-addressed storage to make copies nearly free. When a collection is branched, only the membership list is copied — the actual image files, stored by content hash, are shared between the original and the branch. Changes to one branch (adding, removing, or reordering assets) do not affect the other. Practical applications include client review workflows (branch a collection, remove work-in-progress items, share the branch), version snapshots (freeze a collection at a point in time for archival), and collaborative editing (each team member branches a shared collection and curates independently before merging selections). https://www.numonic.ai/glossary/collection-branching ### Portfolio Distillation The process of progressively filtering a large generative asset library down to a curated portfolio — typically the top 2% of output representing the best work. Uses layered signals including quality scoring, session highlights, and deliberate aesthetic selection. AI art workflows generate far more content than any creator can manually review. A productive ComfyUI session might produce hundreds of images, of which perhaps a dozen are worth keeping and two or three are portfolio-worthy. Portfolio distillation addresses this signal-to-noise challenge through progressive filtering. The distillation process operates in layers. Automated quality scoring eliminates obvious failures (artifacts, blank outputs, corrupted generations). Session-level highlight detection surfaces the statistically unusual outputs within each session. Behavioral signals (which images the creator downloads, shares, or returns to) indicate implicit quality judgments. The final selection remains human — but the preceding layers reduce the decision space from thousands to dozens. https://www.numonic.ai/glossary/portfolio-distillation ### Semantic Search A retrieval method that finds content based on conceptual meaning rather than exact keyword matches, by comparing vector embeddings in high-dimensional space. Enables finding images by visual or conceptual similarity even when query terms share zero keywords with the stored metadata. Semantic search works by converting both the query and the stored content into numerical vectors (embeddings) in the same mathematical space. The search then finds the stored vectors closest to the query vector, as measured by cosine similarity. A query for "cyberpunk street scene" will match an image tagged "neon rain-slicked alley, Blade Runner aesthetic" because the embedding model has learned these concepts are semantically related. For AI asset libraries, semantic search is essential because creators describe their work in natural language that rarely matches the exact vocabulary in generation parameters. However, semantic search alone has limitations — it cannot enforce strict constraints like exact dates or specific tool names. This is why production systems typically combine semantic search with structured filters in a hybrid search architecture. https://www.numonic.ai/glossary/semantic-search ### Temporal Search Querying an asset library using time-based expressions — "what I made last Tuesday," "images from the brutalist architecture session," or "everything from the first three weeks of the project." Treats time periods and creative sessions as first-class navigation dimensions. Temporal search goes beyond simple date filters by understanding creative time in context. A date range filter requires knowing specific dates; temporal search understands relative time ("last week"), session-relative time ("the session where I explored brutalist architecture"), and project-relative time ("early in the project"). The effectiveness of temporal search depends on creative session clustering — without session boundaries, temporal queries can only operate on raw timestamps. With session context, the system can resolve queries like "the session after lunch where I switched to landscape compositions" into a specific set of assets. https://www.numonic.ai/glossary/temporal-search ### Cross-Tool Provenance The challenge of maintaining a continuous provenance chain when creative work flows across multiple AI and traditional tools — from ComfyUI to Photoshop to Midjourney — where each tool has its own metadata format and no awareness of what preceded it. Professional creative workflows are inherently multi-tool. An image might be generated in ComfyUI, refined in Photoshop, used as a reference in Midjourney, and composited again in ComfyUI. At each tool boundary, provenance information is at risk — Photoshop does not read ComfyUI workflow metadata, and Midjourney does not know an image came from Photoshop. Cross-tool provenance is an open architectural problem. Partial solutions include embedding provenance in standards like C2PA that survive across tools, maintaining a centralized provenance registry that tracks assets across tool transitions, and using content hashing to link versions across tool boundaries even when metadata is lost. https://www.numonic.ai/glossary/cross-tool-provenance ### Metadata Stripping The removal of embedded metadata from image files before distribution. Creates a tension for AI-generated content: stripping protects creative process privacy and prevents prompt disclosure, but removes the provenance chain that regulations require and audit trails depend on. Metadata stripping happens both deliberately and accidentally. Social media platforms strip metadata on upload to reduce file sizes and protect user privacy. Creators strip metadata to prevent prompt and workflow disclosure. Image editors strip metadata as a side effect of format conversion. The metadata stripping paradox is that the same metadata serves opposing purposes. Generation parameters are the provenance record that regulators and clients need for compliance verification, but they are also the intellectual property that creators want to protect. Privacy-tiered export — offering different metadata levels for different audiences (full metadata for archive, stripped for social media, selective for client delivery) — is the architectural response to this tension. https://www.numonic.ai/glossary/metadata-stripping ### LoRA (Low-Rank Adaptation) A technique for fine-tuning large AI image generation models on a small dataset to learn a specific style, subject, or concept. LoRA files modify a base model's outputs and require their own metadata management, versioning, and provenance tracking as they directly affect generation character. LoRA works by training a small set of additional weights that are applied on top of a base model like Stable Diffusion XL. Rather than fine-tuning the entire model (which requires significant compute and storage), LoRA modifies only a low-rank subset of the model's attention layers, producing files typically 10-200 MB instead of the base model's 2-7 GB. For asset management, LoRAs introduce unique challenges. They are part of the generation provenance — the same prompt with different LoRAs produces entirely different results. They need version tracking because LoRA files are updated as creators refine their training data. And they need to be linked to the assets they influenced, creating a dependency graph between model components and outputs that traditional DAMs have no concept of. https://www.numonic.ai/glossary/lora ### Model Context Protocol (MCP) An open protocol that allows AI agents and language models to interact with external tools and services through a standardized interface. In creative workflows, MCP enables AI assistants to search, organize, and manage assets in a DAM system directly — making the library agent-accessible. Model Context Protocol defines three core primitives for tool integration: resources (data the tool can share), tools (actions the tool can perform), and prompts (contextual guidance). An asset management system that implements MCP can be accessed by any AI agent that speaks the protocol — from coding assistants to creative workflow automation tools. For creative professionals, MCP integration means an AI assistant can search the asset library for reference images, check for duplicates before uploading, organize new assets into collections, and read metadata — all through natural language requests. The asset library becomes a shared resource accessible to both human users through the UI and AI agents through MCP, enabling workflows where generation tools can reference library assets without manual file management. https://www.numonic.ai/glossary/model-context-protocol ### Agent-First API Design An interface design philosophy where API endpoints are built primarily for machine callers (AI agents) rather than exclusively for human users through a UI. The user interface becomes one consumer among many, changing how operations are scoped, errors are communicated, and tool contracts are specified. Traditional API design optimizes for human developers writing application code. Agent-first API design optimizes for AI agents that discover capabilities at runtime, compose operations dynamically, and handle errors through reasoning rather than hardcoded error handling. Key differences include: operations are discrete and composable rather than bundled into complex flows, error messages are explanatory rather than coded, capabilities are self-describing so agents can discover what actions are available, and responses include context that helps agents decide their next step. The user interface still exists but consumes the same API that agents do, ensuring consistency between human and machine access paths. https://www.numonic.ai/glossary/agent-first-api ### Deduplication The identification and elimination of duplicate assets in a library. Exact deduplication uses content hashing to detect byte-identical files stored under different names. Near-duplicate detection uses visual embedding similarity to identify perceptually identical images that differ at the binary level. AI art libraries accumulate duplicates rapidly. The same image may be saved from the ComfyUI output folder, copied into a project directory, and backed up to cloud storage — three copies with different filenames but identical content. Exact deduplication using content hashes (like SHA-1) catches these perfectly. Near-duplicate detection is more subtle. Two images generated with the same prompt but different seeds may be visually indistinguishable to a human but have completely different binary content. Near-duplicate detection uses embedding similarity — if two images occupy nearly the same position in embedding space, they are flagged as near-duplicates. This helps identify unintentional regenerations and near-identical outputs that clutter the library. https://www.numonic.ai/glossary/deduplication ### Automatic Curation Using quality signals, behavioral patterns, and visual analysis to surface the highest-value assets from a large generative library without manual rating. Signals include technical quality scores, selection frequency, within-session diversity, and aesthetic assessments from embedding-based classifiers. The volume of AI-generated content makes manual curation impractical at scale. A productive creator might generate 500+ images per day, making it impossible to individually review and rate every output. Automatic curation provides a first pass by applying multiple quality signals. Technical quality scoring detects artifacts, blank outputs, and corrupted generations. Behavioral signals — which images the creator downloads, shares, opens repeatedly, or uses as img2img inputs — indicate implicit quality judgments. Visual diversity scoring identifies the most distinctive outputs within a session. These signals combine to create a quality ranking that surfaces the most promising assets for human review, reducing the decision space from hundreds to dozens. https://www.numonic.ai/glossary/automatic-curation ### Style Reference (--sref) Midjourney's --sref parameter applies a consistent visual aesthetic from a reference image or saved style code to new generations. Style references enable brand consistency across prompts by decoupling visual identity from prompt text, allowing teams to enforce approved aesthetics at scale. Style references solve a core governance problem in AI image generation: maintaining visual consistency when multiple team members generate content. Without --sref, even identical prompts produce visually inconsistent results because aesthetic interpretation varies with each generation. A style reference anchors the visual treatment to a known baseline. Professional teams build governed style kits — curated collections of approved --sref codes organized by tier (brand-core, project-specific, experimental). Each reference is tested against standard prompts to document its visual behavior. The challenge is that style references are opaque codes with no inherent description, so without a searchable library linking codes to visual examples, teams lose track of their approved styles within weeks. https://www.numonic.ai/glossary/style-reference ### Variation Lineage The branching history of how an AI-generated image evolved through successive variations, upscales, and remixes from its original generation grid. Lineage tracking preserves the parent-child relationships between outputs so creators can trace any final asset back to its initial prompt and understand the creative decisions at each branch point. In Midjourney, every generation starts as a four-image grid. Selecting an image for variation creates a new branch; upscaling creates a leaf node; remixing creates a branch with a modified prompt. Without lineage tracking, this tree structure exists only in the creator's memory — and is lost entirely when images are downloaded as flat files. For professional workflows, lineage enables three critical capabilities: explaining creative decisions to clients ("we explored this direction but chose this branch because..."), reproducing similar variations for new projects, and auditing the generation process for compliance. A DAM system that captures variation lineage transforms a folder of disconnected images into a navigable creative history. https://www.numonic.ai/glossary/variation-lineage ### Batch Export Downloading multiple AI-generated images simultaneously as a ZIP archive from a platform like Midjourney. Batch exports introduce specific failure modes at scale — ZIP timeouts above 5,000 images, filename collisions causing silent overwrites, and inconsistent metadata embedding — that make post-export verification essential. Batch export is the primary mechanism for moving large image libraries from Midjourney's cloud platform to local or external storage. While single-image downloads are reliable, batch operations at scale encounter server-side timeouts during ZIP packaging, producing incomplete archives with no error notification. Filename collisions are particularly dangerous: when multiple images share generated filenames, extraction silently overwrites earlier files. Professional workflows require a verification layer after every batch export: comparing expected versus received file counts, validating that metadata survived the ZIP packaging, and checking for duplicate filenames before extraction. Without this discipline, teams discover missing assets only when they need them — often months later when the original generation history may no longer be accessible. https://www.numonic.ai/glossary/batch-export ### Prompt Library A structured, searchable collection of generation prompts with their parameters, output examples, and performance history. Unlike simple prompt lists in Notion or spreadsheets, a proper prompt library maintains bidirectional links between prompts and their generated outputs, enabling retrieval by visual result rather than text recall alone. Most creators start with ad-hoc prompt storage — text files, Discord message history, or Notion databases. These approaches break at scale because they treat prompts as isolated text rather than as linked records with context. A well-structured prompt library captures the prompt text, all parameters (--ar, --stylize, --sref, --v), representative output images, usage notes, and performance ratings. The critical capability that separates a prompt library from a prompt list is bidirectional linking: given a prompt, find all images it produced; given an image, find the prompt that created it. This transforms prompt management from "I remember typing something about..." to systematic retrieval. For teams, prompt libraries also enable governance — approved prompts for client work versus experimental prompts for exploration. https://www.numonic.ai/glossary/prompt-library ### Style Drift Gradual visual inconsistency that emerges when multiple team members generate AI images without shared style governance. Style drift occurs when creators use slightly different parameters, unapproved style references, or ad-hoc prompt modifications that individually seem minor but cumulatively produce a fragmented brand aesthetic. Style drift is the organizational equivalent of software entropy — without active governance, visual consistency degrades over time. A team of five creators each making reasonable aesthetic choices will produce noticeably different visual styles within weeks. The problem compounds because drift is difficult to detect incrementally; each individual generation looks acceptable, but the collection lacks coherence. Preventing style drift requires three mechanisms: a governed style kit with approved --sref codes and documented visual behaviors, a review workflow where style choices are validated against brand guidelines before client delivery, and version-controlled style references that track intentional evolution separately from unintended drift. Organizations that treat style governance as optional typically discover the cost during client-facing reviews. https://www.numonic.ai/glossary/style-drift ### trainedAlgorithmicMedia An IPTC Digital Source Type standard value indicating that content was generated by an AI model trained on data. Midjourney embeds this value in downloaded images' IPTC metadata, providing a machine-readable disclosure signal. Unlike C2PA, this tag is unsigned and can be trivially removed, creating a compliance gap for regulations requiring tamper-resistant disclosure. The trainedAlgorithmicMedia value is part of IPTC's Digital Source Type vocabulary, designed to standardize how AI-generated content is labeled in image metadata. When Midjourney embeds this value, it signals to any metadata-aware system that the image was produced by a trained model rather than captured by a camera or created manually. The critical limitation is that IPTC metadata is unsigned — anyone with basic tools can strip or modify it. This matters for EU AI Act Article 50 compliance, which requires AI-generated content to be labeled in a way that is "machine-readable" and resistant to removal. While IPTC trainedAlgorithmicMedia satisfies the machine-readable requirement, it fails the tamper-resistance test. Organizations relying solely on this signal for compliance should understand that downstream recipients can remove the disclosure with no detectable trace. https://www.numonic.ai/glossary/trained-algorithmic-media ### Metadata Recovery The process of reconnecting AI-generated images that have lost their generation metadata to their original prompts, parameters, and provenance records. Recovery techniques include Job ID matching against platform history, visual similarity search against known-good assets, and temporal correlation with generation logs. Metadata recovery addresses the "metadata desert" problem — the millions of AI-generated images downloaded before platforms like Midjourney began embedding generation metadata (pre-late 2025). These legacy files contain no prompt, seed, or parameter information, making them opaque assets with unknown provenance. Recovery is a multi-technique process. The highest-fidelity method is Job ID matching: if the filename or any associated record contains the Midjourney Job ID, the image can be matched against the user's generation history to recover the full prompt and parameters. When Job IDs are unavailable, visual similarity search using perceptual hashing or embedding-based matching can identify near-duplicates that do have metadata. Temporal correlation — matching file timestamps to known generation sessions — provides a lower-confidence but still useful recovery signal. https://www.numonic.ai/glossary/metadata-recovery ### Folder-based Organisation An asset management approach where images are organized primarily by their storage location in a hierarchical folder tree. Folder-based systems force each asset into a single location despite having multiple relevant attributes (project, style, status, date), creating a single-axis retrieval constraint that breaks down above approximately 10,000 assets. Folder-based organisation is the default paradigm inherited from desktop computing. It works well for small collections where the creator remembers what they have and where they put it. The fundamental limitation is that folders impose a single taxonomy: an image can live in "Client-A/Campaign-Q3" or "Portraits/Photorealistic" but not both without duplication. At scale, this single-axis constraint creates three failure modes. First, retrieval degrades because the creator must remember which axis they used when filing. Second, cross-cutting queries become impossible — "show me all photorealistic portraits from Q3 regardless of client" requires manually searching every client folder. Third, governance gaps emerge because folder names carry no machine-readable semantics. Multi-axis organisation through metadata tagging, collections, and search replaces location-based retrieval with attribute-based discovery. https://www.numonic.ai/glossary/folder-based-organisation ### Delivery Package A structured collection of final AI-generated assets prepared for client handoff, including the image files in required formats, usage rights documentation, AI disclosure language for regulatory compliance, generation provenance records, and approval chain documentation linking each deliverable to its approved selection. A delivery package transforms raw AI outputs into a professional deliverable. Beyond the image files themselves, it includes licensing terms specifying permitted uses, AI disclosure statements required by regulations like the EU AI Act or California SB 942, and provenance records linking each final asset to its generation parameters and approval chain. The distinction between sending files and delivering a package is increasingly important as AI content regulations mature. Agencies that deliver images without disclosure documentation transfer compliance liability to clients. A proper delivery package answers every question a client's legal team might ask: What model generated this? Is it disclosed as AI-generated? What are the usage rights? Can we prove the approval chain? Organizations that build delivery packaging into their workflow avoid the retroactive scramble when compliance questions arise. https://www.numonic.ai/glossary/delivery-package ### Provenance Gap A break in the chain of generation metadata that occurs when an AI-generated asset crosses a tool boundary — for example, exporting from Midjourney to Photoshop to ComfyUI. Each tool transition risks losing provenance data because tools use incompatible metadata formats and most editing software discards generation-specific fields on save. Provenance gaps are the multi-tool equivalent of metadata stripping. When an image moves from Midjourney (IPTC/EXIF metadata) to Photoshop (XMP sidecar) to ComfyUI (PNG tEXt chunks), each transition can silently discard the previous tool's provenance data. Photoshop does not understand Midjourney's generation parameters; ComfyUI does not read Photoshop's edit history. The result is an asset whose final state cannot be traced back to its origin. Closing provenance gaps requires an external system of record — a DAM or provenance store — that captures metadata at each tool boundary before the transition occurs. This creates a unified lineage graph that spans tools, rather than relying on each tool to preserve its predecessor's metadata. For compliance purposes, provenance gaps are especially problematic because they can make it impossible to prove an asset's AI-generated origin after post-processing. https://www.numonic.ai/glossary/provenance-gap ### Aesthetic-texture matching Matching input texture richness to the target aesthetic — maximise pores/grain for cinema, restrain for self-tape/phone footage. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Aesthetic-texture matching sits under conditioning and guidance, alongside related concepts such as Classifier-Free Guidance, Dynamic CFG scheduling, and Genetic blueprint. Numonic captures conditioning and guidance like Aesthetic-texture matching as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/aesthetic-texture-matching ### Anti-cut prompting Placing explicit continuity language in the first sentence to suppress the model's default to insert cuts. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Anti-cut prompting sits under prompt craft, alongside related concepts such as Attention budget, Grammatical scoping, and Gravity. Numonic captures prompt craft like Anti-cut prompting as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/anti-cut-prompting ### Model architecture A modern video-generation stack: the diffusion transformer, attention, the VAE, the text encoder, and the dual video/audio streams. Model architecture is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Model architecture as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/architecture ### Asymmetric parameter allocation Dedicating ~14B parameters to the video stream and ~5B to the audio stream, linked by shared timestep conditioning. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Asymmetric parameter allocation sits under model architecture, alongside related concepts such as Attention mechanism, Bidirectional interaction, and Cinematic bias. Numonic captures model architecture like Asymmetric parameter allocation as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/asymmetric-parameter-allocation ### Attention budget The fixed total weight (1.0) the attention mechanism distributes across all tokens in a prompt. Also known as Zero-sum attention. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Attention budget sits under prompt craft, alongside related concepts such as Anti-cut prompting, Grammatical scoping, and Gravity. Numonic captures prompt craft like Attention budget as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/attention-budget ### Attention mechanism The active navigator that aggregates, weights, and shifts across the entire vector field to determine the most probable next state. Also known as Attention. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Attention mechanism sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Bidirectional interaction, and Cinematic bias. Numonic captures model architecture like Attention mechanism as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/attention-mechanism ### Audio-video coupling Unified single-pass generation gives excellent per-take sync but ties vocal identity to each individual seed. Also known as Acoustic coupling. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Audio-video coupling sits under failure mode, alongside related concepts such as Drift inheritance, Environment lock, and Identity popping. Numonic captures failure mode like Audio-video coupling as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/audio-video-coupling ### Base sampler The primary low-resolution DiT stage that builds structure/identity/motion; only ~17% of wall-clock yet does the creative work. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Base sampler sits under sampling and noise schedule, alongside related concepts such as Deep-fried pixels, Denoising process, and Diffusion. It closely corresponds to Sampler in a related medium. Numonic captures sampling and noise schedule like Base sampler as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/base-sampler ### Bidirectional interaction Decoupled streams constantly exchanging information so visual cues map to auditory events with sub-frame precision. Also known as Cross-modal attention. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Bidirectional interaction sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Cinematic bias. Numonic captures model architecture like Bidirectional interaction as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/bidirectional-interaction ### Buffer generation Generating extra seconds so the clean middle can bridge a blend, never inheriting tail-end trash pixels. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Buffer generation sits under production workflow, alongside related concepts such as Glue schedule, Hybrid pipeline, and Video extend. Numonic captures production workflow like Buffer generation as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/buffer-generation ### Center of gravity The coordinate (the weighted mean) where the cumulative pull of all prompt tokens converges; the denoising target. Also known as CoG. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Center of gravity sits under latent geometry, alongside related concepts such as Cross-feeding, Delta vector, and Entropy creep. Numonic captures latent geometry like Center of gravity as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/center-of-gravity ### Chain-of-latents System-2 refinement of complex constraints across the latent path using an inference budget, mimicking human deliberation. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Chain-of-latents sits under inference and reasoning, alongside related concepts such as Chain-of-steps, Chain-of-thought, and Cognitive surrender. Numonic captures inference and reasoning like Chain-of-latents as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/chain-of-latents ### Chain-of-steps The finding that reasoning happens along the latent trajectory, with structural decisions made in specific denoising windows. Also known as CoS. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Chain-of-steps sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-thought, and Cognitive surrender. Numonic captures inference and reasoning like Chain-of-steps as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/chain-of-steps ### Chain-of-thought Forcing calculation of intermediate steps (e.g. how a voice should sound) before rendering the final frame or waveform. Also known as CoT. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Chain-of-thought sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Cognitive surrender. Numonic captures inference and reasoning like Chain-of-thought as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/chain-of-thought ### Cinematic bias The DiT's trained default to introduce dynamic camera moves (pans, tilts, dolly pushes) mimicking cinematic training data. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Cinematic bias sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Cinematic bias as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/cinematic-bias ### Classifier-Free Guidance The inference-time overshoot mechanism extrapolating between conditioned and unconditioned predictions to amplify prompt adherence. Also known as CFG. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Classifier-Free Guidance sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Dynamic CFG scheduling, and Genetic blueprint. It is related to Negative prompt in a related medium. Numonic captures conditioning and guidance like Classifier-Free Guidance as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/classifier-free-guidance ### Close-up Framing tight on a subject's face or a detail. Also known as CU. A form of shot type in cinematography. Within Numonic's Cinematography vocabulary, Close-up sits under shot type. It closely corresponds to Hard SFX in a related medium. Numonic captures shot type like Close-up as structured metadata, so shots, camera moves, and lighting stay searchable and reproducible across AI-generated footage. https://www.numonic.ai/glossary/close-up ### Cognitive surrender The human habit of adopting AI outputs with minimal scrutiny, risking override of human intuition and deliberation. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Cognitive surrender sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like Cognitive surrender as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/cognitive-surrender ### Conditioning and guidance How prompts and reference latents steer the trajectory: CFG, hard/soft conditioning, identity anchoring, state-continuity. Conditioning and guidance is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Conditioning and guidance as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/conditioning ### Context window The 1,000-token prompt memory limit; beyond it, earlier instructions lose their gravity. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Context window sits under embedding and tokens, alongside related concepts such as Embedding, Latent token, and Positional embedding. Numonic captures embedding and tokens like Context window as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/context-window ### Cross-feeding Injecting a percentage of one shot's end-vectors into the next shot's high-noise onset to carry identity, lighting, and camera momentum. Also known as Inertial transfer. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Cross-feeding sits under latent geometry, alongside related concepts such as Center of gravity, Delta vector, and Entropy creep. Numonic captures latent geometry like Cross-feeding as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/cross-feeding ### Decoder Translates the fully denoised latents back into visible pixels and audible sound waves. Also known as VAE decoder. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Decoder sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Decoder as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/decoder ### Decoupled latents Using modality-specific VAEs to compress signals into separate video and audio latents rather than one shared space. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Decoupled latents sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Decoupled latents as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/decoupled-latents ### Deep-fried pixels Neon-glowing edges, crushed blacks, and crunchy texture from high CFG multiplying structural error late in a take. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Deep-fried pixels sits under sampling and noise schedule, alongside related concepts such as Base sampler, Denoising process, and Diffusion. Numonic captures sampling and noise schedule like Deep-fried pixels as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/deep-fried-pixels ### Delta vector The small per-step calculation that subtracts a specific amount of randomness, nudging latents toward a likelier coordinate. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Delta vector sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Entropy creep. Numonic captures latent geometry like Delta vector as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/delta-vector ### Denoising process The iterative method transforming random noise into a structured audiovisual sequence via per-step delta vectors in latent space. Also known as Denoising trajectory. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Denoising process sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Diffusion. Numonic captures sampling and noise schedule like Denoising process as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/denoising-process ### Diffusion Generating structure from chaos by iteratively removing Gaussian noise to reveal an image or sound. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Diffusion sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Diffusion as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/diffusion ### Diffusion Transformer The 22B-parameter asymmetric hybrid marrying transformer scaling with diffusion fidelity; the shot engine's structural backbone. Also known as DiT. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Diffusion Transformer sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. It closely corresponds to Diffusion Transformer in a related medium. Numonic captures model architecture like Diffusion Transformer as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/diffusion-transformer ### Drift Unintended departure of identity/style across frames or versions (the visible symptom of entropy creep). Also known as Latent drift. A form of generation mechanic in cinematography. Within Numonic's Cinematography vocabulary, Drift sits under generation mechanic, alongside related concepts such as Seed. It closely corresponds to Acoustic creep in a related medium. Numonic captures generation mechanic like Drift as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/drift ### Drift inheritance The Extend function treating a drifting tail as the new truth, so the next shot starts broken and collapses faster. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Drift inheritance sits under failure mode, alongside related concepts such as Audio-video coupling, Environment lock, and Identity popping. Numonic captures failure mode like Drift inheritance as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/drift-inheritance ### Dynamic CFG scheduling Front-loading high CFG (7-10) during structural onset to lock identity, then decaying it as entropy creeps in. Also known as CFG decay. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Dynamic CFG scheduling sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Genetic blueprint. Numonic captures conditioning and guidance like Dynamic CFG scheduling as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/dynamic-cfg-scheduling ### Embedding and tokens The numerical representations the model reasons over: text/latent/thinking tokens, positional and rotary embeddings, the context window. Embedding and tokens is a top-level concept in Numonic's Cinematography vocabulary. It is related to Textual inversion in a related medium. Numonic captures cinematography vocabulary such as Embedding and tokens as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/embedding ### Embedding A numerical vector representation of words, images, or audio capturing inherent properties and semantic relationships. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Embedding sits under embedding and tokens, alongside related concepts such as Context window, Latent token, and Positional embedding. Numonic captures embedding and tokens like Embedding as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/embedding-vector ### Entropy A measure of unpredictability/disorder in the latent space; generation moves from maximum (noise) to low (structure). A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Entropy sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Entropy as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/entropy ### Entropy creep The cumulative slow-burn rise of disorder from compounding FP8/BF16 rounding errors that progressively dissolves the structured manifold, producing drift. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Entropy creep sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Entropy creep as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/entropy-creep ### Environment lock Inability to transition between distinct spatial manifolds (interior/exterior, room/room) in a single take without collapse. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Environment lock sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Identity popping. Numonic captures failure mode like Environment lock as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/environment-lock ### Establishing shot Opening wide that situates location/time of a scene. A form of wide shot in cinematography. Within Numonic's Cinematography vocabulary, Establishing shot sits under wide shot. It closely corresponds to Ambience in a related medium. Numonic captures wide shot like Establishing shot as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/establishing-shot ### Failure mode Systemic video-generation failure states and hard walls: text warping, noodle hands, environment lock, drift inheritance, tail-end collapse. Failure mode is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Failure mode as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/failure-mode ### Final descent The last large sigma jump to 0.0 that removes haze and sharpens high-frequency detail (pores, fabric grain, hair edges). A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Final descent sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Final descent as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/final-descent ### Flicker and pop Luminance mismatch between two stitched shots causing the seam to visibly jump; smoothed by the glue schedule. Also known as Flicker. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Flicker and pop sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Flicker and pop as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/flicker-and-pop ### Gaussian noise Random uncorrelated values following the normal distribution; the clean-slate raw material diffusion sculpts. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Gaussian noise sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Gaussian noise as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/gaussian-noise ### Gemma 3 encoder The instruction-tuned text encoder (3,840 hidden dims, 262,208 vocab) parsing prompts into the shared latent manifold. Also known as Gemma 3 12B, or Text encoder. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Gemma 3 encoder sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Gemma 3 encoder as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/gemma3-encoder ### Genetic blueprint Initial latents treated as ground-truth DNA dictating how identity and lighting evolve across the entire sequence. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Genetic blueprint sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Genetic blueprint as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/genetic-blueprint ### Generation mechanic Latent-generation control surface that materially affects matchable output. Generation mechanic is a top-level concept in Numonic's Cinematography vocabulary. It is a broader analogue of Diffusion Transformer in a related medium. Numonic captures cinematography vocabulary such as Generation mechanic as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/gen-mechanic ### Global structure phase The high-sigma early steps that decide global geometry and bone structure rather than colour or texture. Also known as Rough sketch phase. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Global structure phase sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Global structure phase as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/global-structure-phase ### Glue schedule A low-noise refinement schedule (0.25, 0.18, 0.12, 0.06, 0.0) that smooths a blend seam without changing geometry. Also known as Transition sigma schedule. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Glue schedule sits under production workflow, alongside related concepts such as Buffer generation, Hybrid pipeline, and Video extend. Numonic captures production workflow like Glue schedule as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/glue-schedule ### Grammatical scoping Using brackets for global tone and em-dashes to scope cues to specific tokens, preventing attention bleed. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Grammatical scoping sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Gravity. Numonic captures prompt craft like Grammatical scoping as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/grammatical-scoping ### Gravity The attractive pull tokens exert on the trajectory; early prompt content weights more heavily, so anchors go first. Also known as Front-loading. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Gravity sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Gravity as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/gravity ### The great leap The large middle sigma jumps that fast-travel through the manifold once structure is set; over-sampling here breeds drift. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, The great leap sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like The great leap as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/great-leap ### Guidance oversaturation Over-amplifying low-frequency signals via high CFG, producing deep-fried colours and unrealistic contrast. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Guidance oversaturation sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Guidance oversaturation as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/guidance-oversaturation ### Hard-conditioning The starting frame compressed into latents that act as the literal physical denoising state the model extends. Also known as Hard-conditioning latents. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Hard-conditioning sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Hard-conditioning as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/hard-conditioning ### High-noise cluster Tightly clustered early sigma steps (1.0 to 0.975) spending ~50% of compute to lock identity and fight identity popping. Also known as High-noise end. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, High-noise cluster sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like High-noise cluster as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/high-noise-cluster ### Hybrid pipeline Treating the video-generation model as one stage among image-generation, TTS, NLE/VFX, and audio-post tools. Also known as Multi-tool stack. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Hybrid pipeline sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Video extend. Numonic captures production workflow like Hybrid pipeline as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/hybrid-pipeline ### Identity anchor The t=0 frame that establishes and holds subject identity; it degrades as the model applies its transition function over time. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Identity anchor sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Identity anchor as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/identity-anchor ### Identity popping Characters resolving as different people when insufficient high-noise compute fails to lock a stable identity manifold. Also known as Identity pop. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Identity popping sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Identity popping as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/identity-popping ### Inference and reasoning How the model deliberates: System-1 vs System-2 processing, thinking tokens, chains of latents/steps/thought, statistical inference. Inference and reasoning is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Inference and reasoning as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/inference ### Inference budget The compute allotment a System-2 pass spends deliberating over prompt constraints before diffusion begins. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Inference budget sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like Inference budget as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/inference-budget ### Instruction dilution Adding descriptive text reduces adherence to primary intent because every word competes for finite attention weight. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Instruction dilution sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Instruction dilution as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/instruction-dilution ### KISS threshold Pruning to the load-bearing tokens; adherence is ~80% at 3 instructions, ~50% at 8, under 28% at 15. Also known as KISS. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, KISS threshold sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like KISS threshold as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/kiss-threshold ### Laconic prompting Using the minimum words to convey a command, keeping the weighted mean a sharp point rather than a blurry cloud. Also known as Laconic precision. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Laconic prompting sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Laconic prompting as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/laconic-prompting ### Latent blending Mixing the generated latents of two shots inside diffusion (e.g. at sigma 0.5) so features settle on a shared mathematical middle ground. Also known as Latent blend. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Latent blending sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Latent blending as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/latent-blending ### Latent geometry The geometry of the latent space: manifolds, vectors, trajectories, and the forces that pull generation toward a target. Latent geometry is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Latent geometry as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/latent-geometry ### Latent space The high-dimensional space in which the model represents data by essential semantic features rather than raw pixels; similar concepts sit close together. Also known as Compressed latent space. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Latent space sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Latent space as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/latent-space ### Latent token A VAE-produced token representing a patch's textures/colours/edges; the raw diffusion material. Also known as Visual token. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Latent token sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Positional embedding. Numonic captures embedding and tokens like Latent token as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/latent-token ### Latent trajectory The continuous mathematical path the initial latents follow as they transform into generated latents across denoising. Also known as Singular trajectory. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Latent trajectory sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Latent trajectory as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/latent-trajectory ### Latent vector The quantified essence of creative intent — simultaneously a point (state) and an arrow (direction/transformation) in the manifold. Also known as Numerical vector. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Latent vector sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Latent vector as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/latent-vector ### Load-bearing token The 2-4 essential tokens — the unavoidable truth of the shot — that anchor the scene's centre of gravity. Also known as Massive anchor. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Load-bearing token sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Load-bearing token as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/load-bearing-token ### Video extend One-click extension setting the last frame as new hard-conditioning latents; fast but inherits and compounds drift. Also known as Extend. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Video extend sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. It closely corresponds to Audio extend in a related medium. Numonic captures production workflow like Video extend as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/video-extend ### Manifold The topological surface holding the set of all mathematically valid variations of a given subject or environment. Also known as Latent manifold. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Manifold sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Manifold as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/manifold ### Mode collapse The model defaulting to the most generic average interpretation (e.g. 'person walking') when contradictory cues cannot resolve. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Mode collapse sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Mode collapse as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/mode-collapse ### Mode concentration Final-stage CFG amplifying within-mode contraction to lock in fine details and reduce variability. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Mode concentration sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Mode concentration as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/mode-concentration ### Motor-skill failure Chaotic, rapidly occluding hand topologies the 3D temporal RoPE fails to track, so digits merge or multiply. Also known as Noodle hands. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Motor-skill failure sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Motor-skill failure as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/motor-skill-failure ### Multimodal guider A specialised guider splitting CFG into independent text-guidance strength and cross-modal alignment strength. Also known as Multimodal guidance. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Multimodal guider sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Multimodal guider as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/multimodal-guider ### Multimodal integration Treating pixels, audio, and text as numerical vectors in one hidden space, enabling frame-accurate lip-sync. Also known as Multimodal fusion. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Multimodal integration sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Multimodal integration as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/multimodal-integration ### Noise injection Adding timestep-dependent noise directly to the hard-conditioning latents as the raw material the model refines. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Noise injection sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Noise injection as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/noise-injection ### Patch-based processing The VAE dividing a frame into a grid of patches, each compressed into a latent token of textures, colours, and edges. Also known as Patches. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Patch-based processing sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Patch-based processing as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/patch-processing ### Positional embedding Encoding token order/position, since transformers are otherwise position-agnostic about sequence. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Positional embedding sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Positional embedding as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/positional-embedding ### Prompt craft The discipline of writing token-efficient prompts: managing the attention budget, instruction dilution, prosody, scoping, anti-cut language. Prompt craft is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Prompt craft as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/prompt-craft ### Prompt iteration Changing prompt wording, which fundamentally changes the manifold; the fix for blocking, framing, or grammar problems. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Prompt iteration sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. Numonic captures production workflow like Prompt iteration as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/prompt-iteration ### Punctuation as prosody Typographical symbols inside quotes act as literal acoustic tuning: ! shouts, ? lifts, ellipsis decays, dash pauses. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Punctuation as prosody sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. It is related to Vocoder in a related medium. Numonic captures prompt craft like Punctuation as prosody as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/punctuation-as-prosody ### Quantisation Precision tiers (FP16, FP8, GGUF Q8/Q4) trading VRAM for softening that prompts counter with texture cues. Also known as Quantization. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Quantisation sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. Numonic captures production workflow like Quantisation as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/quantisation ### Reachable space The set of mathematically valid variations of a subject's identity established from the initial frame. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Reachable space sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Reachable space as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/reachable-space ### Reasoning High-dimensional statistical inference (geometric optimisation over embeddings), not conscious cognitive judgment. Also known as Latent logic. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Reasoning sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like Reasoning as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/reasoning ### Rotary Positional Embedding Rotates embedding vectors in a complex plane by an angle set by token position, encoding relative distance. Also known as RoPE. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Rotary Positional Embedding sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Rotary Positional Embedding as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/ro-pe ### Sampling and noise schedule The diffusion denoising process and the sigma/noise schedule that allocates compute across the noise range. Sampling and noise schedule is a top-level concept in Numonic's Cinematography vocabulary. It is related to Sampling steps in a related medium. Numonic captures cinematography vocabulary such as Sampling and noise schedule as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/sampling ### Scaling law Transformers handling far more parameters than U-Nets without instability, enabling the asymmetric 22B engine. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Scaling law sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Scaling law as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/scaling-law ### Seed RNG seed; reproducibility anchor and a strong Signal-1 match key. A form of generation mechanic in cinematography. Within Numonic's Cinematography vocabulary, Seed sits under generation mechanic, alongside related concepts such as Drift. It closely corresponds to Seed in a related medium. Numonic captures generation mechanic like Seed as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/seed ### Seed variance Varying the random seed (4-6 takes) to explore different trajectories within the same manifold when performance nuance is off. Also known as Seed rotation. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Seed variance sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. Numonic captures production workflow like Seed variance as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/seed-variance ### Semantic axes Abstract directions in the vector space where each position corresponds to a feature such as warmth, vertical motion, or metallic texture. Also known as Axes of meaning. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Semantic axes sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Semantic axes as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/semantic-axes ### Semantic lock Using high-vector-proximity synonyms (e.g. 'parched' vs 'thirsty') to steer texture toward a neighbourhood without more words. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Semantic lock sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Semantic lock as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/semantic-lock ### Semantic redundancy Synonyms and filler that each consume a token and dilute the gravity of the primary instruction. Also known as Descriptive noise. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Semantic redundancy sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Semantic redundancy as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/semantic-redundancy ### Semantic steering Steering the denoising trajectory toward the prompt's weighted mean using mathematical anchors rather than literal understanding. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Semantic steering sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Semantic steering as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/semantic-steering ### Shared block structure Each transformer block running self-attention, text cross-attention, audio-visual cross-attention, and a feed-forward network in sequence. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Shared block structure sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Shared block structure as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/shared-block-structure ### Sigma The noise level at a given step; 1.0 is total Gaussian noise and 0.0 is a fully denoised image or audio track. Also known as Noise level. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Sigma sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. Numonic captures sampling and noise schedule like Sigma as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/sigma ### Sigma schedule A non-linear sequence of sigmas that reshapes where in the noise range the model spends its compute budget. Also known as ManualSigmas, or Noise schedule. A form of sampling and noise schedule in cinematography. Within Numonic's Cinematography vocabulary, Sigma schedule sits under sampling and noise schedule, alongside related concepts such as Base sampler, Deep-fried pixels, and Denoising process. It closely corresponds to Scheduler in a related medium. Numonic captures sampling and noise schedule like Sigma schedule as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/sigma-schedule ### Soft-conditioning Using an input merely as an inspirational visual reference rather than a literal starting state (contrast hard-conditioning). Also known as Visual reference. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Soft-conditioning sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like Soft-conditioning as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/soft-conditioning ### Spatial compression The VAE's 32x downsampling of physical geometry, why micro-structures like text and hands struggle to survive. Also known as Spatial downsampling. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Spatial compression sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Spatial compression as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/spatial-compression ### Spatial upscaler The heaviest stage (~37% of wall-clock) lifting high-frequency texture from the base sampler's low-resolution foundation. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Spatial upscaler sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. It closely corresponds to Upscaler in a related medium. Numonic captures production workflow like Spatial upscaler as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/spatial-upscaler ### Spurious subtitles Overlaid-subtitle hallucination triggered by mentioning dialogue, captions, or text in the prompt. Also known as Subtitle hallucination. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Spurious subtitles sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Spurious subtitles as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/spurious-subtitles ### State-continuity The architectural rule that the model extends states rather than transforms them, carrying the initial manifold forward. Also known as Inertial force. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, State-continuity sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. Numonic captures conditioning and guidance like State-continuity as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/state-continuity ### Statistical inference Calculating the most probable visual and acoustic patterns for semantic vectors from the training distribution. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Statistical inference sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like Statistical inference as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/statistical-inference ### Stitch and denoiser Generating two shots and blending their overlap with a light denoise pass to purge drift across the seam. Also known as Manual latent blending. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Stitch and denoiser sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. It closely corresponds to Audio stitch and denoiser in a related medium. Numonic captures production workflow like Stitch and denoiser as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/stitch-and-denoiser ### System 1 processing Fast, pattern-based output relying on immediate statistical likelihood; the default mode for most takes. Also known as System 1. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, System 1 processing sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like System 1 processing as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/system1 ### System 2 processing Uses an inference budget and thinking tokens to evaluate complex prompt constraints before the diffusion pass. Also known as System 2. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, System 2 processing sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like System 2 processing as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/system2 ### Tail-end collapse Exponential autoregressive failure at a take's end where limbs morph and identity pops into a new person. Also known as Structural collapse. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Tail-end collapse sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Tail-end collapse as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/tail-end-collapse ### Temporal compression Compressing video across time (8x stride) so one latent token packs several frames; the reason on-screen typography is destroyed. Also known as Temporal downsampling. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Temporal compression sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Temporal compression as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/temporal-compression ### Temporal RoPE Generalises RoPE to 3D spatiotemporal tokens (t,x,y); low-frequency temporal channels avoid phase wrapping over long clips. Also known as 3D RoPE. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Temporal RoPE sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Temporal RoPE as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/temporal-ro-pe ### Temporal text warping 8x temporal compression loses high-frequency edge data, so on-screen letters drift or morph into illegible glyphs. Also known as Text warping. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Temporal text warping sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Temporal text warping as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/temporal-text-warping ### Temporal upscaler Stabilising micro-motion and frame rate after spatial upscaling. A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Temporal upscaler sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. Numonic captures production workflow like Temporal upscaler as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/temporal-upscaler ### Text embedding Gemma-3-encoded vectors capturing prompt semantics that guide the video and audio generation. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Text embedding sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Text embedding as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/text-embedding ### Text token A sub-word unit from the 262,208-token vocabulary; common words are one token, rare words split into several. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Text token sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Text token as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/text-token ### Theatrical projection Over-loud delivery on quiet dialogue caused by a stray exclamation mark or capitals inside the quote. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Theatrical projection sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Theatrical projection as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/theatrical-projection ### Thinking Strategic allocation of compute via thinking tokens to resolve ambiguity, simulating trajectories before committing. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Thinking sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like Thinking as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/thinking ### Thinking token Hidden tokens the encoder generates to deliberate over a prompt before generation, critical for phonetic timing. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Thinking token sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Thinking token as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/thinking-token ### Tiered production architecture Three workflow variants differing in upscaler steps — survey (Rapid), comparison (Fast), shipping (Final). A form of production workflow in cinematography. Within Numonic's Cinematography vocabulary, Tiered production architecture sits under production workflow, alongside related concepts such as Buffer generation, Glue schedule, and Hybrid pipeline. Numonic captures production workflow like Tiered production architecture as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/tiered-production-architecture ### Time budgeting Calibrating dialogue to 2.5-3.0 words per second so the speech fits the shot's duration. Also known as Words per second. A form of prompt craft in cinematography. Within Numonic's Cinematography vocabulary, Time budgeting sits under prompt craft, alongside related concepts such as Anti-cut prompting, Attention budget, and Grammatical scoping. Numonic captures prompt craft like Time budgeting as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/time-budgeting ### Token mass The attention weight/gravity a token carries; load-bearing nouns and verbs are heavy, connector words near-zero. Also known as High-mass token. A form of embedding and tokens in cinematography. Within Numonic's Cinematography vocabulary, Token mass sits under embedding and tokens, alongside related concepts such as Context window, Embedding, and Latent token. Numonic captures embedding and tokens like Token mass as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/token-mass ### Trajectory guidance Point splines drawn on the initial frame infused into latent space to force generated latents along a precise physical path. Also known as Motion tracking, or Point trajectories. A form of conditioning and guidance in cinematography. Within Numonic's Cinematography vocabulary, Trajectory guidance sits under conditioning and guidance, alongside related concepts such as Aesthetic-texture matching, Classifier-Free Guidance, and Dynamic CFG scheduling. It is related to ControlNet in a related medium. Numonic captures conditioning and guidance like Trajectory guidance as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/trajectory-guidance ### Transformer The attention-based architecture processing all tokens simultaneously for global context, scaling hardware-friendly with more parameters. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Transformer sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Transformer as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/transformer ### Variational Autoencoder Compresses raw pixels/audio into latent tokens and decodes them back, using 32x spatial and 8x temporal downsampling. Also known as VAE. A form of model architecture in cinematography. Within Numonic's Cinematography vocabulary, Variational Autoencoder sits under model architecture, alongside related concepts such as Asymmetric parameter allocation, Attention mechanism, and Bidirectional interaction. Numonic captures model architecture like Variational Autoencoder as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/vae ### Vector adjacency Reasoning by computing the mathematical distance between instruction tokens and latent trajectories in hidden space. Also known as Semantic neighbourhood. A form of inference and reasoning in cinematography. Within Numonic's Cinematography vocabulary, Vector adjacency sits under inference and reasoning, alongside related concepts such as Chain-of-latents, Chain-of-steps, and Chain-of-thought. Numonic captures inference and reasoning like Vector adjacency as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/vector-adjacency ### Vector arithmetic Navigating meaning by moving along interpretable latent directions (age, lighting, velocity), inducing semantically coherent transformations. Also known as Semantic traversal. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Vector arithmetic sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Vector arithmetic as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/vector-arithmetic ### Vector leap The mathematical distance moved between latent points; drastic state transitions need large leaps the model suppresses for coherence. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Vector leap sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Vector leap as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/vector-leap ### Vector magnitude The length of a displacement vector, representing the scale of the requested change. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Vector magnitude sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Vector magnitude as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/vector-magnitude ### Voice consistency Vocal identity is generated fresh per seed/prompt and varies across takes — an architectural, not configurational, limit. Also known as Vocal identity. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Voice consistency sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Voice consistency as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/voice-consistency ### Weighted mean The model's most probable statistical interpretation of the whole prompt, used as the navigational target during early denoising. A form of latent geometry in cinematography. Within Numonic's Cinematography vocabulary, Weighted mean sits under latent geometry, alongside related concepts such as Center of gravity, Cross-feeding, and Delta vector. Numonic captures latent geometry like Weighted mean as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/weighted-mean ### Production workflow Production techniques and pipeline stages: extend, stitch-and-denoiser, seed variance, quantisation, tiered architecture, upscalers. Production workflow is a top-level concept in Numonic's Cinematography vocabulary. Numonic captures cinematography vocabulary such as Production workflow as structured metadata, so the look you dialled in on one shot can be found and repeated on the next. https://www.numonic.ai/glossary/workflow ### Workflow archaeology The counter-intuitive truth that base sampler is not the wall-clock bottleneck; upscale and decode dominate — profile first. A form of failure mode in cinematography. Within Numonic's Cinematography vocabulary, Workflow archaeology sits under failure mode, alongside related concepts such as Audio-video coupling, Drift inheritance, and Environment lock. Numonic captures failure mode like Workflow archaeology as structured metadata, so every take stays linked to the prompt and settings that produced it. https://www.numonic.ai/glossary/workflow-archaeology ### Architectural Photography of buildings and structures emphasising line, scale, and perspective. A form of subject in still-image work. Within Numonic's Still Image vocabulary, Architectural sits under subject, alongside related concepts such as Candid, Environmental portrait, and Macro. Numonic captures subject like Architectural as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/architectural ### Aspect ratio The width-to-height proportion of the frame (1:1, 3:2, 16:9). A form of composition in still-image work. Within Numonic's Still Image vocabulary, Aspect ratio sits under composition, alongside related concepts such as Framing devices, Golden ratio, and Headroom. Numonic captures composition like Aspect ratio as structured metadata, so generated images can be tagged, filtered, and re-created with the same parameters. https://www.numonic.ai/glossary/aspect-ratio ### Butterfly lighting A frontal key placed above the lens, casting a butterfly-shaped shadow under the nose. Also known as Paramount lighting. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Butterfly lighting sits under lighting, alongside related concepts such as Chiaroscuro, Color temperature, and Fill light. Numonic captures lighting like Butterfly lighting as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/butterfly ### Candid An unposed, spontaneous capture of a subject. Also known as Reportage. A form of subject in still-image work. Within Numonic's Still Image vocabulary, Candid sits under subject, alongside related concepts such as Architectural, Environmental portrait, and Macro. Numonic captures subject like Candid as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/candid ### Chiaroscuro Strong contrast between light and dark used for dramatic modelling. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Chiaroscuro sits under lighting, alongside related concepts such as Butterfly lighting, Color temperature, and Fill light. It is related to Low-key in a related medium. Numonic captures lighting like Chiaroscuro as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/chiaroscuro ### Color temperature The warmth or coolness of a light source measured in Kelvin. Also known as Kelvin, or White balance. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Color temperature sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Fill light. Numonic captures lighting like Color temperature as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/color-temperature ### ControlNet A conditioning model constraining generation to a pose, depth, edge, or scribble map. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, ControlNet sits under generation parameter, alongside related concepts such as Denoising strength, Inpainting, and Negative prompt. It is related to Trajectory guidance in a related medium. Numonic captures generation parameter like ControlNet as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/control-net ### Denoising strength How much the input image is altered (0 = identical, 1 = ignored) in image-to-image generation. Also known as img2img strength. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Denoising strength sits under generation parameter, alongside related concepts such as ControlNet, Inpainting, and Negative prompt. It closely corresponds to Denoise strength in a related medium. Numonic captures generation parameter like Denoising strength as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/denoise-strength ### Environmental portrait A subject portrayed within their characteristic surroundings to convey context. A form of subject in still-image work. Within Numonic's Still Image vocabulary, Environmental portrait sits under subject, alongside related concepts such as Architectural, Candid, and Macro. Numonic captures subject like Environmental portrait as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/environmental-portrait ### Fill light A secondary, softer light reducing shadow contrast from the key. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Fill light sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. Numonic captures lighting like Fill light as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/fill-light ### Framing devices Foreground elements (doorways, arches, windows) that frame the subject. Also known as Frame within a frame. A form of composition in still-image work. Within Numonic's Still Image vocabulary, Framing devices sits under composition, alongside related concepts such as Aspect ratio, Golden ratio, and Headroom. Numonic captures composition like Framing devices as structured metadata, so generated images can be tagged, filtered, and re-created with the same parameters. https://www.numonic.ai/glossary/framing-devices ### Golden ratio A compositional proportion (~1.618) placing key elements along a spiral or division for natural balance. Also known as Golden spiral. A form of composition in still-image work. Within Numonic's Still Image vocabulary, Golden ratio sits under composition, alongside related concepts such as Aspect ratio, Framing devices, and Headroom. It is related to Framing in a related medium. Numonic captures composition like Golden ratio as structured metadata, so generated images can be tagged, filtered, and re-created with the same parameters. https://www.numonic.ai/glossary/golden-ratio ### Headroom The space between the subject's head and the top edge of the frame. A form of composition in still-image work. Within Numonic's Still Image vocabulary, Headroom sits under composition, alongside related concepts such as Aspect ratio, Framing devices, and Golden ratio. Numonic captures composition like Headroom as structured metadata, so generated images can be tagged, filtered, and re-created with the same parameters. https://www.numonic.ai/glossary/headroom ### Hero still A final, presentation-grade single image. A form of still role in still-image work. Within Numonic's Still Image vocabulary, Hero still sits under still role, alongside related concepts such as 3D-bound still. Numonic captures still role like Hero still as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/hero-still ### Inpainting Regenerating a masked region of an image while preserving the rest. Also known as Masked generation. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Inpainting sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Negative prompt. Numonic captures generation parameter like Inpainting as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/inpainting ### Key light The primary, dominant light source shaping the subject. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Key light sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. Numonic captures lighting like Key light as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/key-light ### Leading room Space left in front of a subject's gaze or direction of motion. Also known as Nose room. A form of composition in still-image work. Within Numonic's Still Image vocabulary, Leading room sits under composition, alongside related concepts such as Aspect ratio, Framing devices, and Golden ratio. Numonic captures composition like Leading room as structured metadata, so generated images can be tagged, filtered, and re-created with the same parameters. https://www.numonic.ai/glossary/leading-room ### Light quality Small-source hard light gives sharp shadows; large-source soft light gives gradual shadows. Also known as Hard light, or Soft light. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Light quality sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. Numonic captures lighting like Light quality as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/light-quality ### Loop lighting A key slightly off-axis casting a small loop-shaped shadow from the nose. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Loop lighting sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. Numonic captures lighting like Loop lighting as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/loop-lighting ### Macro Extreme close-range photography rendering small subjects at life-size or greater. A form of subject in still-image work. Within Numonic's Still Image vocabulary, Macro sits under subject, alongside related concepts such as Architectural, Candid, and Environmental portrait. Numonic captures subject like Macro as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/macro ### Negative prompt Text describing what to exclude, steering generation away from unwanted features. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Negative prompt sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Inpainting. It is related to Classifier-Free Guidance in a related medium. Numonic captures generation parameter like Negative prompt as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/negative-prompt ### Practical light A light source visible within the frame (lamp, window) that motivates the scene's lighting. Also known as Motivated light. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Practical light sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. It is related to Diegetic in a related medium. Numonic captures lighting like Practical light as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/practical-light ### Product shot A controlled studio capture of a product as the hero subject. Also known as Packshot. A form of subject in still-image work. Within Numonic's Still Image vocabulary, Product shot sits under subject, alongside related concepts such as Architectural, Candid, and Environmental portrait. Numonic captures subject like Product shot as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/product-shot ### Rembrandt lighting A portrait key creating a small triangle of light on the shadowed cheek. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Rembrandt lighting sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. Numonic captures lighting like Rembrandt lighting as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/rembrandt ### Sampler The numerical solver that denoises the diffusion latent at each step. Also known as Sampling method. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Sampler sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Inpainting. It closely corresponds to Base sampler in a related medium. Numonic captures generation parameter like Sampler as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/sampler ### Scheduler The function setting the noise level at each step. Also known as Karras. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Scheduler sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Inpainting. It closely corresponds to Sigma schedule in a related medium. Numonic captures generation parameter like Scheduler as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/scheduler ### Split lighting A key at ~90 degrees lighting exactly half the face while the other half stays in shadow. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Split lighting sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. Numonic captures lighting like Split lighting as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/split-lighting ### Sampling steps The number of denoising iterations; more steps refine detail at higher compute cost. Also known as Steps. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Sampling steps sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Inpainting. It is related to Sampling and noise schedule in a related medium. Numonic captures generation parameter like Sampling steps as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/steps ### Textual inversion A learned embedding token capturing a concept or style, invoked by keyword. Also known as TI. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Textual inversion sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Inpainting. It is related to Embedding and tokens in a related medium. Numonic captures generation parameter like Textual inversion as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/textual-inversion ### 3D-bound still A still produced as an input/reference toward a 3D asset, not a final. A form of still role in still-image work. Within Numonic's Still Image vocabulary, 3D-bound still sits under still role, alongside related concepts such as Hero still. It is related to Large Reconstruction Model in a related medium. It is related to Geometry in a related medium. Numonic captures still role like 3D-bound still as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/three-d-bound-still ### Three-point lighting The standard key + fill + back/rim lighting arrangement. A form of lighting in still-image work. Within Numonic's Still Image vocabulary, Three-point lighting sits under lighting, alongside related concepts such as Butterfly lighting, Chiaroscuro, and Color temperature. It closely corresponds to Lighting in a related medium. Numonic captures lighting like Three-point lighting as structured metadata, so a promising frame keeps the seed and settings behind it instead of becoming a one-off. https://www.numonic.ai/glossary/three-point-lighting ### Upscaler A model that increases resolution and adds high-frequency detail. Also known as Hires fix. A form of generation parameter in still-image work. Within Numonic's Still Image vocabulary, Upscaler sits under generation parameter, alongside related concepts such as ControlNet, Denoising strength, and Inpainting. It closely corresponds to Spatial upscaler in a related medium. Numonic captures generation parameter like Upscaler as structured metadata, so an image you liked months ago is one search away, parameters intact. https://www.numonic.ai/glossary/upscaler ### Vignette Darkening (or lightening) toward the frame edges that draws the eye inward. A form of composition in still-image work. Within Numonic's Still Image vocabulary, Vignette sits under composition, alongside related concepts such as Aspect ratio, Framing devices, and Golden ratio. Numonic captures composition like Vignette as structured metadata, so generated images can be tagged, filtered, and re-created with the same parameters. https://www.numonic.ai/glossary/vignette ### Ambient occlusion A baked map darkening contact creases and cavities to fake soft contact shadowing. Also known as AO. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, Ambient occlusion sits under surfacing, alongside related concepts such as Texture baking, Decal, and Displacement. Numonic captures surfacing like Ambient occlusion as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/ambient-occlusion ### Background removal Isolating the subject on a solid or transparent background before upload; busy or subject-coloured backgrounds corrupt reconstruction. Also known as Cutout. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Background removal sits under input-image craft, alongside related concepts such as 3D-safe framing, Deep focus, and Flat lighting. Numonic captures input-image craft like Background removal as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/background-removal ### Texture baking Transferring high-poly detail (normals / AO / curvature) into maps for a low-poly target. Also known as Bake, or Baking. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, Texture baking sits under surfacing, alongside related concepts such as Ambient occlusion, Decal, and Displacement. Numonic captures surfacing like Texture baking as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/bake ### Baked-shadow dent A hard cast shadow in the input sculpted into the mesh as a physical dent; cheapest to prevent, brutal to fix. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Baked-shadow dent sits under 3d failure mode, alongside related concepts such as Confabulated geometry, Hair lumping, and Hollow fill. Numonic captures 3d failure mode like Baked-shadow dent as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/baked-shadow-dent ### Blockout A rough, untextured massing pass that establishes scale and layout before detailing. Also known as Graybox, or Greybox. A form of geometry in 3D work. Within Numonic's 3D Composition vocabulary, Blockout sits under geometry, alongside related concepts such as Mesh, Point cloud, and Gaussian splat. Numonic captures geometry like Blockout as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/blockout ### 3D-safe framing The rule that a 3D-input still should be deliberately un-cinematic — flat light, deep focus, subject square-on and whole, plain keyable background. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, 3D-safe framing sits under input-image craft, alongside related concepts such as Background removal, Deep focus, and Flat lighting. Numonic captures input-image craft like 3D-safe framing as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/bridge-rule ### Confabulated geometry Hidden, concave, or back-facing geometry the model never saw and invents, usually wrongly. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Confabulated geometry sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Hair lumping, and Hollow fill. Numonic captures 3d failure mode like Confabulated geometry as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/confabulated-geometry ### Decal A crisp logo or text reprojected as a texture overlay rather than modelled as geometry. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, Decal sits under surfacing, alongside related concepts such as Ambient occlusion, Texture baking, and Displacement. Numonic captures surfacing like Decal as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/decal ### Decimation Reducing polygon count of a dense mesh while preserving silhouette, for real-time or LOD use. Also known as Decimate. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Decimation sits under production hand-off, alongside related concepts such as Game-ready mesh, Level of detail, and Poly budget. Numonic captures production hand-off like Decimation as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/decimation ### Deep focus Everything-sharp input focus (the opposite of shallow depth of field), preserving the edge cues the reconstructor needs. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Deep focus sits under input-image craft, alongside related concepts such as Background removal, 3D-safe framing, and Flat lighting. Numonic captures input-image craft like Deep focus as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/deep-focus ### Displacement A map that physically offsets surface geometry — true depth rather than faked normal detail. Also known as Heightmap. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, Displacement sits under surfacing, alongside related concepts such as Ambient occlusion, Texture baking, and Decal. Numonic captures surfacing like Displacement as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/displacement ### Diffusion Transformer A transformer-backbone diffusion model; a common backbone for 3D generation. Also known as DiT. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Diffusion Transformer sits under 3d generation method, alongside related concepts such as Image-to-3D, Large Reconstruction Model, and Multi-view diffusion. It is a broader analogue of Generation mechanic in a related medium. It closely corresponds to Diffusion Transformer in a related medium. Numonic captures 3d generation method like Diffusion Transformer as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/di-t ### Flat lighting Soft, even, shadowless illumination on the input, so cast shadows are not sculpted into the mesh as dents. Also known as Shadowless. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Flat lighting sits under input-image craft, alongside related concepts such as Background removal, 3D-safe framing, and Deep focus. Numonic captures input-image craft like Flat lighting as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/flat-lighting ### Game-ready mesh A clean-topology, low-poly, UV'd mesh that drops into an engine with minimal or no retopology (e.g. image-to-3D tool output). Also known as Gameready. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Game-ready mesh sits under production hand-off, alongside related concepts such as Decimation, Level of detail, and Poly budget. Numonic captures production hand-off like Game-ready mesh as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/game-ready ### 3D generation method The generative method that produced a 3D asset — matchable provenance for a generated model. 3D generation method is a top-level concept in Numonic's 3D Composition vocabulary. Numonic captures 3D vocabulary such as 3D generation method as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/gen-method ### Geometry How a 3D asset's form is represented: polygon mesh, point cloud, Gaussian splat, or rough blockout. Geometry is a top-level concept in Numonic's 3D Composition vocabulary. It is related to 3D-bound still in a related medium. Numonic captures 3D vocabulary such as Geometry as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/geometry ### Hair lumping Hair and fur reconstructed as solid bulk mass rather than strands or cards; hero work grooms hair natively instead. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Hair lumping sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hollow fill. Numonic captures 3d failure mode like Hair lumping as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/hair-lumping ### Production hand-off Turning a raw generated sculpt into a production-usable asset: retopology, UVs, PBR, interchange, and provenance. Production hand-off is a top-level concept in Numonic's 3D Composition vocabulary. Numonic captures 3D vocabulary such as Production hand-off as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/handoff ### High-poly A dense, high-detail mesh — a digital sculpt or the source before decimation. Also known as Highpoly, or Sculpt. A form of mesh in 3D work. Within Numonic's 3D Composition vocabulary, High-poly sits under mesh, alongside related concepts such as Low-poly. Numonic captures mesh like High-poly as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/high-poly ### Hollow fill The model filling openings and cavities that should stay hollow, producing solid where the subject was open. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Hollow fill sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Hollow fill as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/hollow-fill ### Image-to-3D Generating a 3D asset from a single reference image. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Image-to-3D sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Large Reconstruction Model, and Multi-view diffusion. Numonic captures 3d generation method like Image-to-3D as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/image-to3-d ### Input-image craft How to frame the source still for image-to-3D: the reconstructor imagines geometry from pixels, so ambiguity becomes hallucinated geometry. Input-image craft is a top-level concept in Numonic's 3D Composition vocabulary. Numonic captures 3D vocabulary such as Input-image craft as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/input-craft ### Level of detail A set of progressively simpler mesh versions swapped by distance to keep a real-time scene performant. Also known as LOD. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Level of detail sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Poly budget. Numonic captures production hand-off like Level of detail as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/lod ### Low-poly A mesh with a deliberately small polygon count, for real-time or stylised use. Also known as Lowpoly. A form of mesh in 3D work. Within Numonic's 3D Composition vocabulary, Low-poly sits under mesh, alongside related concepts such as High-poly. Numonic captures mesh like Low-poly as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/low-poly ### Large Reconstruction Model A feed-forward model that reconstructs 3D geometry from one or a few images in seconds. Also known as LRM. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Large Reconstruction Model sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Multi-view diffusion. It is related to 3D-bound still in a related medium. Numonic captures 3d generation method like Large Reconstruction Model as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/lrm ### Mesh Polygonal surface representation made of vertices, edges, and faces. Also known as Polygons, or Polymesh. A form of geometry in 3D work. Within Numonic's 3D Composition vocabulary, Mesh sits under geometry, alongside related concepts such as Blockout, Point cloud, and Gaussian splat. Numonic captures geometry like Mesh as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/mesh ### Multi-view diffusion Generating several view-consistent 2D images with a diffusion model, then reconstructing 3D from them; risks view disagreement / ghosting. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Multi-view diffusion sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Multi-view diffusion as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/multi-view-diffusion ### Multiview-to-3D Generating a 3D asset from several overlapping views of one subject; more depth cues than a single image. Also known as Multiview. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Multiview-to-3D sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Multiview-to-3D as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/multiview-to3-d ### 3D-native latent diffusion Diffusing directly in a 3D latent space (the prior is genuinely 3D, not 2D-projected), yielding cleaner watertight geometry. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, 3D-native latent diffusion sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like 3D-native latent diffusion as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/native3-d-diffusion ### Neural radiance field An implicit radiance field learned from posed images for novel-view synthesis. Also known as NeRF. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Neural radiance field sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Neural radiance field as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/ne-rf ### Non-manifold geometry Edges shared by more than two faces or otherwise unclean topology that breaks downstream tools until repaired. Also known as Non-manifold. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Non-manifold geometry sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Non-manifold geometry as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/non-manifold ### Normal map A tangent-space map that fakes high-frequency surface detail on a low-poly mesh. Also known as Normalmap, or Normals. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, Normal map sits under surfacing, alongside related concepts such as Ambient occlusion, Texture baking, and Decal. Numonic captures surfacing like Normal map as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/normal-map ### Object fusion Multiple objects in one input merging into a single fused mass; avoided by one-subject-per-run. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Object fusion sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Object fusion as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/object-fusion ### Orthographic turnaround A consistent front/side/back/top image set that lets the model recover otherwise-hidden geometry. Also known as Turnaround. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Orthographic turnaround sits under input-image craft, alongside related concepts such as Background removal, 3D-safe framing, and Deep focus. Numonic captures input-image craft like Orthographic turnaround as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/orthographic-turnaround ### Over-densification A dense, evenly-triangulated sculpt with no clean edge flow; the reason hero output must be retopologised. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Over-densification sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Over-densification as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/over-densification ### PBR materials Physically based materials (albedo / metallic / roughness / normal) that light consistently across scenes. Also known as PBR. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, PBR materials sits under surfacing, alongside related concepts such as Ambient occlusion, Texture baking, and Decal. Numonic captures surfacing like PBR materials as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/pbr ### Photogrammetry Reconstructing 3D geometry and texture from overlapping photographs. Also known as Photoscan. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Photogrammetry sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Photogrammetry as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/photogrammetry ### Point cloud An unstructured set of 3D points, typically from scanning, prior to meshing. Also known as Pointcloud. A form of geometry in 3D work. Within Numonic's 3D Composition vocabulary, Point cloud sits under geometry, alongside related concepts such as Blockout, Mesh, and Gaussian splat. Numonic captures geometry like Point cloud as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/point-cloud ### Poly budget The target polygon / face count an asset class is allowed, set by its role (hero vs background prop). Also known as Facelimit. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Poly budget sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Level of detail. Numonic captures production hand-off like Poly budget as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/poly-budget ### Asset provenance The recorded chain from source still → model version → retopo → licence tier for a generated 3D asset. Also known as Lineage. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Asset provenance sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Level of detail. Numonic captures production hand-off like Asset provenance as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/provenance ### Quad remesh Rebuilding a triangulated sculpt as an even quad mesh with better edge flow for deformation. Also known as Quadremesh, or Remesh. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Quad remesh sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Level of detail. Numonic captures production hand-off like Quad remesh as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/quad-remesh ### Rectified flow A flow-matching generative formulation (e.g. a rectified-flow 3D generator) that learns near-straight transport paths for fast, stable 3D shape generation. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Rectified flow sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Rectified flow as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/rectified-flow ### Retopology Rebuilding a dense or scanned mesh as clean, animation-ready topology. Also known as Retopo. A form of topology in 3D work. Within Numonic's 3D Composition vocabulary, Retopology sits under topology. Numonic captures topology like Retopology as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/retopo ### Rig The skeletal / control structure that lets a 3D model be posed or animated. Also known as Armature, Rigged, or Skeleton. Rig is a top-level concept in Numonic's 3D Composition vocabulary. Numonic captures 3D vocabulary such as Rig as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/rig ### Riggable pose An A-pose or T-pose with limbs separated from the torso, so the mesh does not fuse arm-to-body and can be rigged. Also known as A-pose, or T-pose. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Riggable pose sits under input-image craft, alongside related concepts such as Background removal, 3D-safe framing, and Deep focus. Numonic captures input-image craft like Riggable pose as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/riggable-pose ### Scale and orientation drift Generated meshes carry no inherent real-world scale or up-axis; both must be standardised on ingest. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Scale and orientation drift sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Scale and orientation drift as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/scale-drift ### Software-defined asset The MovieLabs 2030 direction where an asset is a metadata-rich, provenance-carrying package rather than a bare mesh file. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Software-defined asset sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Level of detail. Numonic captures production hand-off like Software-defined asset as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/software-defined-asset ### Gaussian splat A radiance-field scene represented as oriented 3D Gaussians, for real-time novel-view rendering. Also known as 3DGS, Splat, or Splatting. A form of geometry in 3D work. Within Numonic's 3D Composition vocabulary, Gaussian splat sits under geometry, alongside related concepts such as Blockout, Mesh, and Point cloud. Numonic captures geometry like Gaussian splat as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/splat ### Subject isolation One subject per generation; multi-object inputs fuse, so separate runs are reassembled in a DCC. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Subject isolation sits under input-image craft, alongside related concepts such as Background removal, 3D-safe framing, and Deep focus. Numonic captures input-image craft like Subject isolation as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/subject-isolation ### Surfacing How a 3D surface looks: UV layout, physically based materials, textures, and detail maps. Surfacing is a top-level concept in Numonic's 3D Composition vocabulary. Numonic captures 3D vocabulary such as Surfacing as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/surfacing ### Text melting On-surface text and logos smearing into illegible geometry; treated as a reprojected decal, not modelled as shape. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Text melting sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Text melting as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/text-melting ### Text-to-3D Generating a 3D asset from a text prompt. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Text-to-3D sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Text-to-3D as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/text-to3-d ### Texture An image map applied to a surface — the base colour / albedo and its companion maps. Also known as Albedo, or Basecolor. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, Texture sits under surfacing, alongside related concepts such as Ambient occlusion, Texture baking, and Decal. Numonic captures surfacing like Texture as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/texture ### Three-quarter view A front-plus-one-side hero view that yields the most depth from a single input image. A form of input-image craft in 3D work. Within Numonic's 3D Composition vocabulary, Three-quarter view sits under input-image craft, alongside related concepts such as Background removal, 3D-safe framing, and Deep focus. Numonic captures input-image craft like Three-quarter view as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/three-quarter-view ### Topology The edge/face layout of a mesh; clean topology deforms and textures predictably. Also known as Edgeflow. A form of geometry in 3D work. Within Numonic's 3D Composition vocabulary, Topology sits under geometry, alongside related concepts such as Blockout, Mesh, and Point cloud. Numonic captures geometry like Topology as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/topology ### Transparency lumping Glass, liquid, and other refractive surfaces sculpted as solid opaque lumps rather than thin-walled transparent shells. A form of 3d failure mode in 3D work. Within Numonic's 3D Composition vocabulary, Transparency lumping sits under 3d failure mode, alongside related concepts such as Baked-shadow dent, Confabulated geometry, and Hair lumping. Numonic captures 3d failure mode like Transparency lumping as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/transparency-lumping ### Triplane A compact 3D representation storing features on three orthogonal feature planes; the output of many large reconstruction models. Also known as Tri-plane. A form of 3d generation method in 3D work. Within Numonic's 3D Composition vocabulary, Triplane sits under 3d generation method, alongside related concepts such as Diffusion Transformer, Image-to-3D, and Large Reconstruction Model. Numonic captures 3d generation method like Triplane as structured metadata, so meshes, textures, and their source stills stay linked and reproducible through the generation pipeline. https://www.numonic.ai/glossary/triplane ### OpenUSD The Universal Scene Description interchange standard; layering and metadata carry an asset's lineage as first-class data. Also known as USD, or USDZ. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, OpenUSD sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Level of detail. Numonic captures production hand-off like OpenUSD as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/usd ### UV mapping The 2D parameterisation that maps textures onto a 3D surface. Also known as UV, UVs, or Unwrap. A form of surfacing in 3D work. Within Numonic's 3D Composition vocabulary, UV mapping sits under surfacing, alongside related concepts such as Ambient occlusion, Texture baking, and Decal. Numonic captures surfacing like UV mapping as structured metadata, so a finished asset can be traced back to the reference image and settings that made it. https://www.numonic.ai/glossary/uv ### Watertight geometry A closed, manifold mesh with no holes — required for boolean ops, 3D printing, and clean simulation. Also known as Watertight. A form of production hand-off in 3D work. Within Numonic's 3D Composition vocabulary, Watertight geometry sits under production hand-off, alongside related concepts such as Decimation, Game-ready mesh, and Level of detail. Numonic captures production hand-off like Watertight geometry as structured metadata, so the pipeline from source still to game-ready mesh stays searchable end to end. https://www.numonic.ai/glossary/watertight ### Acoustic creep Voice turning metallic and prosody degrading over successive extends as the phonetic manifold is lost. Also known as Audio latent drift. A form of audio continuity in audio production. Within Numonic's Audio Production vocabulary, Acoustic creep sits under audio continuity, alongside related concepts such as Audio extend, and Audio stitch and denoiser. It closely corresponds to Drift in a related medium. Numonic captures audio continuity like Acoustic creep as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/acoustic-creep ### Ambience The continuous background environmental sound bed that establishes place. Also known as Atmos, or Room tone. A form of stem in audio production. Within Numonic's Audio Production vocabulary, Ambience sits under stem, alongside related concepts such as Dialogue, Foley, and Music. It closely corresponds to Establishing shot in a related medium. It is related to Wide shot in a related medium. Numonic captures stem like Ambience as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/ambience ### Audio extend The model carrying the acoustic state (room tone, mid-sentence) into a continuation; risks acoustic creep. Also known as Acoustic relay. A form of audio continuity in audio production. Within Numonic's Audio Production vocabulary, Audio extend sits under audio continuity, alongside related concepts such as Acoustic creep, and Audio stitch and denoiser. It closely corresponds to Video extend in a related medium. Numonic captures audio continuity like Audio extend as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/audio-extend ### Audio stitch and denoiser Latent-blending audio vectors for a phase-aligned cross-fade that prevents the DC-offset click of a hard cut. Also known as Audio latent blend. A form of audio continuity in audio production. Within Numonic's Audio Production vocabulary, Audio stitch and denoiser sits under audio continuity, alongside related concepts such as Acoustic creep, and Audio extend. It closely corresponds to Stitch and denoiser in a related medium. Numonic captures audio continuity like Audio stitch and denoiser as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/audio-stitch-denoiser ### Background The most distant sonic layer (ambience, walla) sitting behind the mix. A form of audio plane in audio production. Within Numonic's Audio Production vocabulary, Background sits under audio plane, alongside related concepts such as Foreground, and Midground. It is related to Low-key in a related medium. Numonic captures audio plane like Background as structured metadata, so dialogue, music, and sound design stay searchable alongside the footage they accompany. https://www.numonic.ai/glossary/background ### Bed A sustained underlying layer of music or ambience over which foreground elements sit. Also known as Ambient bed, or Music bed. A form of audio form in audio production. Within Numonic's Audio Production vocabulary, Bed sits under audio form, alongside related concepts such as Cue, and Transition. Numonic captures audio form like Bed as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/bed ### Audio continuity Carrying the acoustic state across generated segments without an audible seam or identity break. Audio continuity is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as Audio continuity as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/continuity ### Cue A discrete bounded segment of music or sound scored to a specific moment or scene. Also known as Music cue. A form of audio form in audio production. Within Numonic's Audio Production vocabulary, Cue sits under audio form, alongside related concepts such as Bed, and Transition. Numonic captures audio form like Cue as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/cue ### Designed SFX A synthesised or layered non-literal effect created for impact (whoosh, riser, sci-fi tone). Also known as Sound design. A form of sfx kind in audio production. Within Numonic's Audio Production vocabulary, Designed SFX sits under sfx kind, alongside related concepts such as Hard SFX. Numonic captures sfx kind like Designed SFX as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/designed-sfx ### Dialogue The recorded or generated spoken lines of on-screen or off-screen characters. Also known as DX, Dialog, VO, Voice-over, or Voiceover. A form of stem in audio production. Within Numonic's Audio Production vocabulary, Dialogue sits under stem, alongside related concepts such as Ambience, Foley, and Music. Numonic captures stem like Dialogue as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/dialogue ### Diegesis Whether a sound originates inside or outside the story world. Diegesis is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as Diegesis as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/diegesis ### Diegetic Sound whose source exists within the story world and can be heard by the characters. Also known as Source audio. A form of diegesis in audio production. Within Numonic's Audio Production vocabulary, Diegetic sits under diegesis, alongside related concepts such as Non-diegetic. It is related to Practical light in a related medium. Numonic captures diegesis like Diegetic as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/diegetic ### Foley Performed everyday sounds (footsteps, cloth, prop handling) recorded in sync to picture. A form of stem in audio production. Within Numonic's Audio Production vocabulary, Foley sits under stem, alongside related concepts such as Ambience, Dialogue, and Music. It closely corresponds to Extreme close-up in a related medium. Numonic captures stem like Foley as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/foley ### Foreground The closest, most prominent sonic layer, usually dialogue or a key effect. A form of audio plane in audio production. Within Numonic's Audio Production vocabulary, Foreground sits under audio plane, alongside related concepts such as Background, and Midground. It is related to High-key in a related medium. Numonic captures audio plane like Foreground as structured metadata, so dialogue, music, and sound design stay searchable alongside the footage they accompany. https://www.numonic.ai/glossary/foreground ### Audio form A bounded unit of audio content: a cue, a bed, or a transition between sections. Audio form is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as Audio form as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/form ### Hard SFX A specific sync-to-picture effect tied to a visible on-screen action (door slam, gunshot). Also known as Spot effect. A form of sfx kind in audio production. Within Numonic's Audio Production vocabulary, Hard SFX sits under sfx kind, alongside related concepts such as Designed SFX. It closely corresponds to Close-up in a related medium. Numonic captures sfx kind like Hard SFX as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/hard-sfx ### Midground The intermediate sonic layer sitting between foreground performance and background ambience. A form of audio plane in audio production. Within Numonic's Audio Production vocabulary, Midground sits under audio plane, alongside related concepts such as Background, and Foreground. Numonic captures audio plane like Midground as structured metadata, so dialogue, music, and sound design stay searchable alongside the footage they accompany. https://www.numonic.ai/glossary/midground ### Music Composed musical material scoring a scene, either diegetic or non-diegetic. Also known as Score. A form of stem in audio production. Within Numonic's Audio Production vocabulary, Music sits under stem, alongside related concepts such as Ambience, Dialogue, and Foley. It is related to Camera movement in a related medium. Numonic captures stem like Music as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/music ### Non-diegetic Sound outside the story world (score, narration) that the characters cannot hear. A form of diegesis in audio production. Within Numonic's Audio Production vocabulary, Non-diegetic sits under diegesis, alongside related concepts such as Diegetic. Numonic captures diegesis like Non-diegetic as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/non-diegetic ### Phonetic timing The natural rhythm/timing of speech the model computes via thinking tokens before rendering the waveform. Also known as Prosody. A form of audio synthesis in audio production. Within Numonic's Audio Production vocabulary, Phonetic timing sits under audio synthesis, alongside related concepts such as Vocoder. Numonic captures audio synthesis like Phonetic timing as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/phonetic-timing ### Audio plane The depth layer a sound occupies in the mix (foreground, midground, background). Audio plane is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as Audio plane as structured metadata, so dialogue, music, and sound design stay searchable alongside the footage they accompany. https://www.numonic.ai/glossary/plane ### Point of audition The listening perspective the mix adopts — whose ears we hear from; the audio analogue of point of view. Also known as POA. Point of audition is a top-level concept in Numonic's Audio Production vocabulary. It closely corresponds to Point-of-view shot in a related medium. Numonic captures audio-production vocabulary such as Point of audition as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/point-of-audition ### Sound effects Non-speech, non-music sounds representing the actions and events in a scene. Also known as SFX. A form of stem in audio production. Within Numonic's Audio Production vocabulary, Sound effects sits under stem, alongside related concepts such as Ambience, Dialogue, and Foley. Numonic captures stem like Sound effects as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/sfx ### SFX kind Whether an effect is a literal sync-to-picture hard effect or a synthesised designed effect. SFX kind is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as SFX kind as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/sfx-kind ### Silence The deliberate absence of sound used for dramatic or rhythmic effect (cf. the field manual's 'allowable silence' in the speech budget). A form of stem in audio production. Within Numonic's Audio Production vocabulary, Silence sits under stem, alongside related concepts such as Ambience, Dialogue, and Foley. It is related to Negative space in a related medium. Numonic captures stem like Silence as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/silence ### Stem A grouped audio mixdown of one category (dialogue, music, or effects) kept separate for mixing and delivery. Also known as Submix. Stem is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as Stem as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/stem ### Audio synthesis How the model renders speech and sound: computing phonetic timing then translating it to waveforms. Audio synthesis is a top-level concept in Numonic's Audio Production vocabulary. Numonic captures audio-production vocabulary such as Audio synthesis as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/synthesis ### Transition A sonic bridge (crossfade, segue, stinger) connecting two sections. Also known as Crossfade, or Segue. A form of audio form in audio production. Within Numonic's Audio Production vocabulary, Transition sits under audio form, alongside related concepts such as Bed, and Cue. Numonic captures audio form like Transition as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/transition ### Vocoder The component matching typographical prosody triggers (! ? ... dash) to specific frequency and amplitude waveforms. A form of audio synthesis in audio production. Within Numonic's Audio Production vocabulary, Vocoder sits under audio synthesis, alongside related concepts such as Phonetic timing. It is related to Punctuation as prosody in a related medium. Numonic captures audio synthesis like Vocoder as structured metadata, so sound work stays organised next to the shots it belongs to. https://www.numonic.ai/glossary/vocoder ### Walla Indistinct background crowd murmur or chatter suggesting a populated space. Also known as Crowd. A form of stem in audio production. Within Numonic's Audio Production vocabulary, Walla sits under stem, alongside related concepts such as Ambience, Dialogue, and Foley. Numonic captures stem like Walla as structured metadata, so a take keeps its settings and stays matched to the picture it was cut against. https://www.numonic.ai/glossary/walla ## Blog - [The New Numonic: Built for Libraries That Got Big](https://www.numonic.ai/blog/new-numonic-app-large-libraries): The redesigned Numonic app is now the default for every account. Here is what changed for people whose AI image libraries outgrew scrolling and browsing. - [Video, Audio, and 3D Deserve Real Previews](https://www.numonic.ai/blog/video-audio-3d-asset-previews): AI output stopped being only images. Numonic generates real thumbnails and previews for video, audio, and 3D assets, however they arrived in your library. - [Bulk Importing AI Images Without Losing Your Place](https://www.numonic.ai/blog/bulk-import-ai-images-without-duplicates): A large import that fails halfway leaves you guessing what landed. Smart Import resumes where it stopped, retries failed files, and creates no duplicates. - [Your Old Midjourney Images Still Have Their Prompts](https://www.numonic.ai/blog/recover-midjourney-prompts-old-images): Midjourney writes your prompt into the image and into its filename. Numonic now surfaces those prompts for images already in your library, even older uploads. - [AI Storage Doesn't Die, It Compresses and Gets Model-Coupled](https://www.numonic.ai/blog/ai-storage-compresses-not-disappears): A new 2.07-billion-request study on AI image storage settles the "will regeneration replace storage" debate. The answer changes what teams should actually optimize for. - [How to Audit an AI Model's Self-Reports](https://www.numonic.ai/blog/audit-ai-model-self-reports): Three techniques for testing what a language model actually knows, none of which use its own testimony as the measurement. Repeatable in an afternoon. - [AI Copyright Rulings Test One Thing: Can You Prove Your Contribution](https://www.numonic.ai/blog/ai-copyright-rulings-burden-of-proof-lineage): German courts ruled on AI-generated art and copyright. The real question they are asking is not "was AI used" — it is whether you can reconstruct what you did. - [What a Frontier Model Can't See About Itself](https://www.numonic.ai/blog/frontier-model-self-knowledge-limits): Three layers of machine self-access: what a model reports, what it represents but withholds, and what leaves no trace at all. They need different fixes. - [The Six-Month Gap in a Model's Knowledge Cutoff](https://www.numonic.ai/blog/claude-opus-5-knowledge-cutoff-gap): A stated knowledge cutoff describes the training corpus, not what the model can recall. I measured the gap on Claude Opus 5. It was six months wide. - [We Made Our AI Coding Agent Judge Its Own Cost](https://www.numonic.ai/blog/ai-agent-judge-its-own-cost): We kept overspending by running our best AI coding agent at maximum effort on routine tasks. Here's the self-judgment pattern that fixed it. - [AI 2040’s Missing Layer: Content Provenance in the Basin of Sanity](https://www.numonic.ai/blog/ai-2040-plan-a-missing-content-layer): AI 2040 designs epistemic infrastructure but skips the content layer. Why provenance is the missing piece—and what the EU has enforced since August 2, 2026. - [The AI Multiplier Isn’t the Model: It’s the Harness](https://www.numonic.ai/blog/ai-multiplier-harness-not-model): The teams compounding fastest with AI aren’t the ones waiting for the next model. They’re building the scaffolding that turns each release into shipped work. - [The Babbage Trap: When to Stop Mining for Gold Because You Suspect There’s a Path to Diamonds](https://www.numonic.ai/blog/babbage-trap-strategic-pivots-ai): A founder’s framework for strategic pivots in the AI age, drawn from Charles Babbage, the Kelly Criterion, and the cautionary mirror of HMS Victoria. When the diamond seam is real but the gold mine is not yet finished. - [How We Used zlib Compression Ratios to Stop NotebookLM from Drowning in Codebase Noise](https://www.numonic.ai/blog/notebooklm-context-drowning-zlib-fix): NotebookLM ignored half our codebase. A zlib compression heuristic and XML-bound atomic packs turned a blind LLM into an expert architectural reviewer. - [EU AI Act Article 50: What Tools Can Actually Do for Your Studio (and What Is Just Marketing)](https://www.numonic.ai/blog/eu-ai-act-article-50-mandatory-vs-marketing): Article 50 of the EU AI Act has been enforceable since August 2, 2026. Here is how to read a vendor compliance page like a regulator would, the three terms vendors blur, and an honest map of what Numonic can and cannot do today. - [Casting Shakespeare With AI: A Public R&D Diary](https://www.numonic.ai/blog/studio-memory-shakespeare-thought-experiment): Why studio memory matters more than generation. Notes from a CTO and CEO running a public AI-video R&D experiment with a four-actor pillar troupe. - [The Creative Agentic Stack: Six Layers for AI-Native Media Infrastructure](https://www.numonic.ai/blog/creative-agentic-stack): The generic agent stack misses three layers creative teams actually need. Where we think AI-native media infrastructure has to go next, and an invitation to AI video and film teams building through it with us. - [EU AI Act Article 50: A CTO’s Reading at T-100](https://www.numonic.ai/blog/eu-ai-act-article-50-t-minus-100): Article 50 of the EU AI Act became enforceable on August 2, 2026, and has applied since. A T-100 look back at what was settled, what was still in negotiation, and what creative studios needed to do before the deadline. - [GPT Image 2 in ComfyUI: Reasoning-Driven Generation, Real Cost Data, and Why Every Output Needs Provenance](https://www.numonic.ai/blog/gpt-image-2-comfyui-complete-guide): How to use GPT Image 2 in ComfyUI via Comfy Cloud Partner Nodes — real credit/timing data, dense-text rendering, edit fidelity, and full-metadata capture. - [Automate the Boring Parts of Your Creative Workflow](https://www.numonic.ai/blog/pipeline-builder-creative-automation): Build asset processing pipelines visually or as text. Twenty-nine stage types, eight templates, cron scheduling, and event triggers for AI creators. - [Workflow Time Travel: See How Your ComfyUI Parameters Evolved](https://www.numonic.ai/blog/workflow-time-travel-comfyui): Numonic fingerprints your ComfyUI workflows and shows how parameters drifted across iterations. Stop guessing what changed between run 30 and run 45. - [Your Prompts Deserve a Library, Not a Clipboard](https://www.numonic.ai/blog/prompt-library-ai-optimizer): Stop losing your best prompts. Numonic extracts every prompt from your AI-generated assets into a searchable, filterable library with thumbnails and version history. - [The Three Layers of Content Trust: Identity, Creation, and Distribution](https://www.numonic.ai/blog/three-layers-content-trust): Content provenance needs three layers working together: who published it (identity), how it was made (creation), and where it travels (distribution). Most solutions only cover one. - [AI as a Governance Partner: A Year of Building with Claude Code](https://www.numonic.ai/blog/ai-governance-partner-claude-code): Three years of AI-assisted development, from chat interfaces to governed in-codebase partnership. Here is what happens when you treat AI as a governance partner, not just a code generator. - [New Feature: Asset Inspection Lightbox with 800% Zoom and Comparison Modes](https://www.numonic.ai/blog/asset-inspection-lightbox): Inspect AI-generated images at full resolution without downloading. Zoom to 800%, compare variations with synced viewports, and review metadata in a fullscreen lightbox. - [The Agency Revenue Case for AI Governance](https://www.numonic.ai/blog/ai-governance-agency-revenue-case): AI governance is a profit center, not a cost center. Data on 2-3x pricing premiums, 2.3x client lifetime value, and 85-89% efficiency gains for governed agencies. - [New Feature: Sessions View with AI Creative Reflections](https://www.numonic.ai/blog/sessions-by-creative-session): Numonic now groups your AI-generated images by creative session and lets you generate per-session AI reflections. No folders. No tagging. Just your creative flow, understood. - [From Chaos to Catalog: How Numonic Auto-Builds Your Model Library from ComfyUI Workflows](https://www.numonic.ai/blog/from-chaos-to-catalog-auto-model-library): Stop manually tracking LoRAs and checkpoints. Learn how Numonic auto-detects models from ComfyUI workflows and builds a browsable, searchable library. - [The Agentic Creative Workflow Is Here. Is Your Asset Management Ready?](https://www.numonic.ai/blog/agentic-creative-workflow-asset-management): AI agents now orchestrate ComfyUI, Stable Diffusion, and Midjourney from the terminal. One of them can batch-generate hundreds of images in minutes. Here is what changes and how to build infrastructure that keeps up. - [App Mode Is ComfyUI's Platform Moment—and a Signal for Infrastructure](https://www.numonic.ai/blog/app-mode-is-comfyui-s-platform-moment-and-a-signal-for-infrastructure): ComfyUI App Mode transforms a node editor into a distribution platform. We analyze what this means for the generative AI stack and the infrastructure layer beneath it. - [Bulk Export Failure Modes and Defences: Why Your Midjourney Archive Breaks](https://www.numonic.ai/blog/midjourney-export-bulk-reliability): ZIP timeouts, incomplete downloads, and filename collisions quietly corrupt Midjourney archives. Here is a reliability playbook for bulk exports at scale. - [Organise Midjourney Images After Export: Inside Midjourney vs Everywhere Else](https://www.numonic.ai/blog/midjourney-10000-images-organisation): Midjourney's web app organises images inside the platform. But creative work happens outside it. Here is how to bridge the export gap at 5,000+ images. - [Your Midjourney Prompt Library Has Outgrown Notion](https://www.numonic.ai/blog/midjourney-prompt-library-beyond-notion): Notion is where prompt libraries start. But prompts without visual context, asset linking, or version lineage are just text. Here is what a governed prompt library looks like. - [Save Midjourney Prompts Like a Pro: From Snippets to Managed IP](https://www.numonic.ai/blog/midjourney-save-prompts-pro): Most MJ users save prompts in notes or spreadsheets. But prompts are now embedded in your downloads — and they travel with every file you share. Here is how to manage prompts as IP. - [Finding That One Midjourney Image From Three Months Ago](https://www.numonic.ai/blog/midjourney-prompt-search-history): You remember the idea but not the prompt, the folder, or the date. MJ's search works inside the platform — but your exported files are unsearchable. Here is how to close the retrieval gap. - [The Agency Guide to Midjourney: From Brief to Deliverable](https://www.numonic.ai/blog/midjourney-agency-workflow): Midjourney is excellent for concepting. But without governance, agencies lose iterations, leak prompts to clients, and have no audit trail. Here is a 6-phase production workflow. - [AI Disclosure for Midjourney Exports: What Embedded Metadata Is Not, and What Compliance Looks Like](https://www.numonic.ai/blog/midjourney-eu-ai-act-compliance): Midjourney's IPTC Digital Source Type tag is a real standard — but unsigned metadata is a statement, not a proof. Here is what Article 50 has required since 2 August 2026. - [Safe Sharing: When to Strip Midjourney Metadata, How to Keep Internal Provenance](https://www.numonic.ai/blog/midjourney-privacy-sharing-guide): Downloaded MJ PNGs embed your full prompt in metadata. Share an image, share your creative process. Here is a redact-vs-preserve decision framework. - [The Midjourney Delivery Package: Hand Off AI Images Without Losing Context — or Oversharing It](https://www.numonic.ai/blog/midjourney-client-delivery-package): Clients need high-res finals and AI disclosure. They should not see your prompts, iterations, or MJ username. Here is a standardised delivery checklist. - [Building a Searchable Style Reference Library Beyond SrefHunt](https://www.numonic.ai/blog/midjourney-style-reference-library): SrefHunt and Midlibrary are great for discovering styles. But discovery is not governance. Here is how to build a style library with approval status, usage tracking, and export. (97 more posts at https://www.numonic.ai/blog) ## ComfyUI node reference - [KSampler](https://www.numonic.ai/integrations/comfyui/ksampler): The primary sampling node in ComfyUI. Runs the diffusion process that transforms random noise into a coherent image using a model, positive/negative conditioning, and configurable parameters like seed, steps, CFG scale, and sampler algorithm. - [KSampler (Advanced)](https://www.numonic.ai/integrations/comfyui/ksampler-advanced): An extended version of KSampler with fine-grained control over the noise addition and start/end steps. Used for advanced workflows like multi-pass rendering, img2img refinement, and ControlNet-guided generation. - [Load Checkpoint](https://www.numonic.ai/integrations/comfyui/checkpoint-loader-simple): Loads a Stable Diffusion checkpoint model file (.safetensors or .ckpt) and outputs the MODEL, CLIP, and VAE components. The primary entry point for model loading in most ComfyUI workflows. - [Load LoRA](https://www.numonic.ai/integrations/comfyui/lora-loader): Applies a LoRA (Low-Rank Adaptation) weight file to an existing model and CLIP encoder. LoRAs modify the base model behavior—adding styles, concepts, or characters—without replacing the full checkpoint. - [Load VAE](https://www.numonic.ai/integrations/comfyui/vae-loader): Loads a standalone VAE (Variational Autoencoder) model file to override the VAE bundled with a checkpoint. Used when a higher-quality or specialized VAE produces better image decoding results. - [DualCLIPLoader](https://www.numonic.ai/integrations/comfyui/dual-clip-loader): Loads two CLIP text encoder models simultaneously, required for architectures like SDXL and Flux that use dual text encoders (CLIP-L and CLIP-G or T5) for richer prompt understanding. - [CLIP Text Encode (Prompt)](https://www.numonic.ai/integrations/comfyui/clip-text-encode): Converts a text prompt into CONDITIONING tokens that guide the diffusion sampling process. This is where the user prompt enters the generation pipeline, making it the primary source of creative intent metadata. - [Load ControlNet Model](https://www.numonic.ai/integrations/comfyui/control-net-loader): Loads a ControlNet model that enables image-guided generation. ControlNets add spatial control (edges, depth, pose, segmentation) to the diffusion process, allowing precise composition from reference images. - [Apply ControlNet](https://www.numonic.ai/integrations/comfyui/control-net-apply): Applies a loaded ControlNet model to conditioning with a reference image and strength parameter. This node is where spatial guidance from a reference image merges into the generation pipeline. - [VAE Decode](https://www.numonic.ai/integrations/comfyui/vae-decode): Decodes a latent representation into a visible pixel image using a VAE (Variational Autoencoder). This is the final conversion step that transforms the sampler output into a viewable image. - [VAE Encode](https://www.numonic.ai/integrations/comfyui/vae-encode): Encodes a pixel image into a latent representation using a VAE. Used in img2img workflows where an existing image serves as the starting point for generation instead of random noise. - [Upscale Image](https://www.numonic.ai/integrations/comfyui/image-scale): Resizes an image to specified dimensions using configurable interpolation methods (nearest, bilinear, bicubic, lanczos). Used for resolution changes, upscaling, and preparing images for further processing. - [Upscale Latent](https://www.numonic.ai/integrations/comfyui/latent-upscale): Resizes a latent representation before decoding, enabling hi-res fix workflows that sample at a higher resolution in a second pass. Operates in latent space for better quality than pixel-space upscaling. - [Save Image](https://www.numonic.ai/integrations/comfyui/save-image): The primary output node in ComfyUI. Writes generated images to disk as PNG files and embeds the full workflow JSON in PNG metadata chunks, preserving the complete generation graph for reproducibility. - [Load Image](https://www.numonic.ai/integrations/comfyui/load-image): Loads an image from the ComfyUI input directory for use in img2img, ControlNet, or other image-based workflows. Records only the filename, creating a fragile provenance link to the source material. - [Empty Latent Image](https://www.numonic.ai/integrations/comfyui/empty-latent-image): Creates a blank latent tensor of specified dimensions, serving as the canvas for txt2img generation. The width and height set here determine the output image resolution (in latent space, 8× for pixels). - [Save Video](https://www.numonic.ai/integrations/comfyui/save-video): ComfyUI's built-in video-output node. Saves a VIDEO input to MP4, MKV, or WebM using H.264 or AV1, and embeds the prompt and workflow as container metadata — no custom node pack or FFmpeg binary required. - [Video Combine (VHS)](https://www.numonic.ai/integrations/comfyui/vhs-video-combine): Combines a batch of frames into a video file — the standard video-output node for ComfyUI animation workflows. Choose video/h264-mp4 with yuv420p for the broadest playback compatibility; WebM, AV1, ProRes, FFV1, GIF, and WebP cover web, archival, and preview needs. - [Load Video (VHS)](https://www.numonic.ai/integrations/comfyui/vhs-load-video): Loads a video file and extracts individual frames as an image batch for processing in ComfyUI. Part of the VideoHelperSuite extension, enabling video-to-video workflows. - [AnimateDiff Loader](https://www.numonic.ai/integrations/comfyui/animate-diff-loader): Loads an AnimateDiff motion model and injects it into the base diffusion model, enabling frame-coherent animation generation. AnimateDiff adds temporal awareness to standard image generation models. - [AnimateDiff Combine](https://www.numonic.ai/integrations/comfyui/animate-diff-combine): Combines AnimateDiff-generated frames into a video or GIF output. The animation-specific alternative to VHS_VideoCombine, with settings tailored for AnimateDiff frame sequences. - [SaveVideo vs VHS_VideoCombine](https://www.numonic.ai/integrations/comfyui/compare/savevideo-vs-vhs-videocombine): Compare ComfyUI's built-in SaveVideo with VideoHelperSuite's VHS_VideoCombine: inputs, outputs, formats, audio, and how each embeds workflow metadata.