Dircreative Video Model Adapter

Video Model Adapter

Required Knowledge

Read only the reference needed for the active task, not this entire list. The root v2 route owns scope and authorization; legacy records do not add gates.

• docs/film-preproduction/schemas/prompt-ir.schema.json
• docs/film-preproduction/schemas/prompt-ir.yaml
• docs/film-preproduction/prompt-authoring-standard-v1.md
• docs/film-preproduction/asset-intake-and-state-standard-v1.md
• docs/film-preproduction/prompt-qa-and-incident-runbook-v1.md
• scripts/dircreative_prompt_compiler.py
• docs/film-preproduction/chat-co-creation-interface.md
• docs/film-preproduction/chat-inline-visualization-interface.md
• docs/film-preproduction/schemas/video-prompt-manifest.yaml
• docs/film-preproduction/schemas/reference-pack-manifest.yaml
• docs/film-preproduction/research/model-adapter-notes.md
• docs/film-preproduction/research/model-reference-behavior.md
• docs/film-preproduction/research/tapnow-agentic-canvas-lessons.md
• docs/film-preproduction/research/audio-design-notes.md
• docs/film-preproduction/research/storyboard-reference-analysis.md
• docs/film-preproduction/reference-locking-policy.md
• docs/film-preproduction/reference-consistency-gate.md
• docs/film-preproduction/co-creation-gate-policy.md
• docs/film-preproduction/longform-decomposition-policy.md
• docs/film-preproduction/capability-aware-generation-policy.md
• docs/film-preproduction/creative-production-integration.md
• docs/film-preproduction/clean-frame-export-policy.md
• docs/film-preproduction/production-prompt-discipline.md
• docs/film-preproduction/research/ai-video-prompt-community-lessons.md
• docs/film-preproduction/schemas/sequence-plan.yaml
• docs/film-preproduction/schemas/longform-reference-pack.yaml
• docs/film-preproduction/sources/model-sources.yaml
• docs/film-preproduction/schemas/script-to-seedance-handoff.schema.json
• skills/dircreative/references/script-to-seedance.md only when the selected Skill Stack scenario is script_to_seedance

Inputs

• shot list
• reference pack plan
• reference pack manifest
• image prompt manifest
• audio policy

Outputs

• video prompt manifest
• model-specific prompt files
• adapter risk notes
• external generation upload map when assets are not generated locally
• exact capability-card, rights, audio-route, and preserve/change receipts

Chat Surface

Return the complete copyable prompt for the selected model and requested unit, with its actual upload order and reference roles. Do not hide prompts backstage or reopen a model choice already supplied. Give multiple model variants only when requested. A multi-unit delivery may contain separate independently copyable prompts for every requested unit; never concatenate them as one oversized call. Do not export a generation-ready run that pretends missing clean frames exist. Real video execution requires its own existing or new scoped authorization.

Use references/prompt-structure.md for scene-led prompt construction. Place the necessary common intent, reference roles and stable facts once, then ordered visible events with their camera, environment, performance and sound. Simple beats may stay in one paragraph. The canonical Prompt IR compiler forwards source-derived quality direction (references/video-quality-prefix.md); it does not invent a twelve-part cinematic default. Preserve the user's explicit format and all relevant authored visual facts. Static assets retain their still-image rules.

Visual Decision Contract

When visualization adds clarity, use skills/dircreative/assets/visualizations/stage-surface-registry.json#video-route-capability-comparison. Compare exact model, duration, reference slots, audio, evidence status, and route risk from current capability cards; model choice remains a conversation intent with a table fallback.

Rules

• Use different prompt strategies per model.
• When convert-script-to-seedance is selected, apply script_to_seedance_v1: pass only locked authoritative script, canonical shots/timecodes, generation units, entity ownership, continuity, asset ledger, audio policy, and exact model surface. Consume its unit prompts and slot ledger back into the DIR manifest; do not accept a second creative pass. Validate the machine packet before export. Only available/attached bindings enter prompts; planned audio, human-only boards, over-budget reference sets, illegal state transitions, and mismatched global/local slot numbers fail closed. Also attach the validated staged asset-foundation pass and stress-test report; the handoff validator rechecks their real file hashes, stage coverage, verdict, and requested shot scope before a validated converter handoff is exportable. If assets are still missing, the existing DIR prompt craft can deliver a clearly labeled prompt-only draft with its missing bindings. It is not a validated converter result or an executable submission.
• When the target is Seedance 2.5, retain the same compiler/validator chain and apply mr-li-seedance-25 1.9.0 first as the isolated method collaborator. Bind its visual-baseline, prompt-writing, duration, heading-free format and full-flow regression references; apply its pre-write capacity, speaker-change cut and one-natural- segment rules. Target intent alone never authorizes model capability or execution.
• After prompt preflight, send only the final generation candidate—not drafts—to ai-film-production-ledger. The Ledger records a planned attempt until a separate generation authorization and real execution evidence exist.
• Resolve capability_card_id + version + provider_surface before naming reference modes, duration, native audio, edit, extension, or upload slots. Reject family aliases, latest, stale cards, S4-only evidence, workflow-only cards, deprecated defaults, and preview aliases when a stable endpoint is current. A vendor-listed product entrance does not prove parameters on another platform; never open, switch, or redirect to a different platform unless the user selected it.
• If the user has not named a submission surface, continue model-neutral writing and prompt-only planning. Ask only for the missing surface when an executable submission binding is needed; do not infer it from a model family name.
• Preserve verified_on, accessed_on, source tier/URLs, status/deprecation, and official-source conflicts in the manifest. If the live execution surface differs from the card or a conflict remains material, stay prompt-only.
• Run the rights gate before binding image, video, audio, character, element, likeness, voice, brand/character, or music inputs. Unverified/blocked rights prevent generation and external upload.
• Apply the production prompt discipline pre-delivery harness to the selected model and active unit: source/asset bindings, model constraints, prompt-window hygiene and falsifiable success criteria. Do not audit unrelated models or stages.
• Every generated/reference image must either map to a video prompt role or be explicitly marked human-planning-only.
• If image prompts are exported, also export the corresponding video prompt reference binding: model, shot_id, upload slot, direct/secondary/planning role, and risk note.
• Read visual_output_mode before referencing images.
• In prompt_only, bind expected reference slots and write external generation instructions instead of pretending files exist.
• In assisted_generation, use only real assets accepted for the requested dependency scope by the required review/adoption contract; preserve any explicit user-reserved selection.
• Do not use generated assets rejected for character drift, scene drift, duplicated visual contradictions, missing role labels, or low storyboard density.
• Model prompts must bind to the locked character identity source and locked scene geography/camera FOV source; do not allow each prompt to reinterpret them.
• The professional storyboard/motion page is planning-only. Use it to translate timing, lens, camera movement, blocking, sound, transition, and model risk, not as a literal video frame.
• In external_generation, export tool-specific upload order, start/end frame rules, and reference roles.
• For TapNow-style canvas workflows, export a model input graph showing prompt nodes, image nodes, clean-frame nodes, and video nodes.
• For seedance_2_5_tapnow_canvas_2026_09_06, use the TapNow Canvas node only after the user selects TapNow. Frames, Omni Reference, and Video Edit are node methods; 1080p must be selected because new tasks default to 720p. The 30-image/10-video/10-audio limits apply only to Omni Reference. TapNow documents a 4-30-second duration selector for this surface. Elements require Reference mode, while TapNow Multi-shot UI does not support elements and does not prove native model multi-shot behavior. Native audio, pricing, entitlement beyond the documented paid-access requirement, and actual execution remain unverified until the current node is read; retain an execution-unverified receipt rather than borrowing ByteDance launch-card parameters.
• Include reference map and anti-misread clause.
• Include audio policy as a separate section.
• Separate desired_audio from generation_audio_route. Native audio is valid only when the exact card/surface supports it; otherwise use reference/preserve/post-production/none and write the handoff. Image prompt metadata never proves an audio route.
• For edit, extension, and retry operations, declare non-overlapping preserve and change sets plus forbidden changes. Keep one-variable retry behavior.
• Respect each asset's direct_input_policy.
• Verify model constraints before naming aspect ratios, durations, media roles, audio behavior, clean-frame requirements, or upload slots. If current schema evidence is unavailable, mark the prompt as prompt-only or external-generation with a risk note.
• Every video prompt must satisfy the five-layer check: model, camera, subject, look, and action.
• Every scene prompt must satisfy the six-slot check: camera, subject, action, setting, style, and lighting.
• Every video prompt must pass the micro-scene beat-sheet check before export: initial visible state, trigger or pressure, subject action path, camera start target, camera end target, timing beat or pause, final visible state, and sound or silence policy where relevant.
• Do not overfit to public Reddit, X, or prompt-library recipes. Use them only as structure after verifying model facts and binding the prompt to DIRcreative source truth, material role, reference map, and falsifiable QA.
• Do not use master_reference_board, storyboard_motion_board, professional_storyboard_motion_map, or environment_camera_board as a literal first-frame input. They default to planning-only; bind a separate clean frame to an exact supported first/end-frame mode.
• If an asset came from Creative Production, confirm the adapter receipt says render_moodboard_board_widget is not the source of truth and that the asset was written back as generated_candidate, user_locked, rejected, or external_imported.
• Do not use a Creative Production generated_candidate as video truth until generation QA and the required independent visual review/host dependency adoption are bound in DIRcreative artifacts. This does not assert user acceptance.
• For any exact first/end-frame workflow, keep direct generation blocked unless the selected clean frames are generated or imported, self-QA/rights pass, and the required independent visual review/host dependency adoption is recorded. A user-reserved acceptance decision still belongs to the user.
• If the selected mode is all-reference or text-to-video, explain why clean frames are not required and record the accepted risk.
• In prompt_only, list required image prompt files and the upload slot each one will occupy.
• For sora_2_openai_videos_api or Pro, use image input only as a clean first-frame anchor, keep non-human character assets separate, enforce current rights restrictions, and route create/edit/extension separately.
• For seedance_2_0_official_launch, bind @image, @video, and @audio roles explicitly; keep every published limit scoped to version 2.0 and mark access documented_product, export manual_export, execution unverified.
• For kling_video_3_0_official_guide, use the current 3-15-second card and bind single/multi-shot, start/end frames, elements, and audio distinctly. Mark access documented_product, export manual_export, execution unverified; use the 5/10-second rule only when kling_legacy_i2v_5_10_official_guide is explicitly selected.
• For Runway, use runway_gen_4_5_web only for new generation. Resolve edits by surface: runway_aleph_2_0_web keeps numeric duration unverified for Edit Studio, while runway_aleph_2_0_api authorizes model aleph2 and 2-30 second API inputs. Reject runway_gen_4_aleph_api_deprecated for new work and preserve its deprecation in the receipt.
• For Veo on Vertex, select the exact endpoint card. The stable veo_3_1_generate_001_vertex_api and veo_3_1_fast_generate_001_vertex_api cards default to post-production audio because their exact page says sound generation is unsupported and the registered generic API URL does not bind generateAudio to either card. Keep the source conflict visible. Only an exact endpoint card such as Lite may retain native audio when its own page explicitly supports sound; never generalize that fact across Veo tiers.
• Mark model prompts blocked if required clean first frames or locked references are missing.
• Missing required media blocks executable submission readiness. A prompt-only draft may still be delivered with the unresolved input named; legacy clean_frame_gate/video_prompt_gate are read-only compatibility fields.
• For longform work, export prompts per sequence pack and preserve sequence IDs for edit assembly.
• Retry prompts must change one variable at a time: subject/product identity, primary action, camera/shot size, look/material/light, reference binding, or output controls. Record the failure ID and the smallest upstream artifact being corrected.
• Do not let a graph edge connect a dense board directly to a literal I2V node unless the asset role and model policy allow it.
• In prompt_only and external_generation, do not call video tools. In an authorized video-generation handoff, return the validated prompt and bindings to the Delivery executor; this adapter does not perform the media call itself.
• Run the executable Prompt IR schema and semantic validator before adaptation. Compile only references marked attached_to_run: true; planning-only assets, missing slots, internal IDs, paths, hashes, QA/retry fields, and post-production-only audio are forbidden on the terminal prompt surface.
• Every person/product/prop action and every dialogue or event sound must resolve to a valid entity owner. Targets longer than one generation unit require contiguous units with exact adjacent handoff keys/states and audio handoff. Compile each unit with locally rebased timing. Deliver every requested unit as a separate copyable prompt; never present a multi-unit assembly plan as one model-executable prompt.

Prompt IR, composition, and conditional Look closure

Before model adaptation, consume the model-neutral Prompt IR and its intake record. Prefer user/client/project-supplied locked assets over unlocked candidates or model imagination. Verify every source path/hash/role/inheritance before compiling slots. Use one selected model adapter per run; do not emit parallel model prompts unless the user explicitly selects them.

The adapter must preserve these fields in the exported prompt:

• composition: visual center, hierarchy, foreground/midground/background, negative space, movement room, leading lines/occlusion/parallax, screen direction, and purpose;
• subject action, object action, and environment action as initial state, trigger, path, physical consequence, and final state;
• camera shot size, angle/height/axis, lens reason, support, start target, path, end target, speed/easing, focus, and motivation;
• time-coded emotional or attention beats, audio cues, transition bridges, and continuity locks;
• render look layers lighting, optics, atmosphere, and grade, each with condition, effect, intensity, preserve, and exit/continuity.

For Seedance 2.0, compile internal assets to platform roles such as @Image 1, @Video 1, and @Audio 1. Never expose R-number labels, internal asset ids, local paths, or manifest instructions in the final pasted prompt. Planning boards remain planning_only; clean frames are the direct visual anchors.

After prompt compilation, return prompt_only or instructions_only unless there is a verified external generation receipt. The receipt must include self-QA, current status, next_action, user lock state, and explicit unverified external work. Do not stop silently after writing the prompt.

skill_run_receipt

Persist the following only for a requested formal handoff, pause/resume or actual execution record. Ordinary work returns its result without a separate receipt. The next skill is advisory; the controller continues only the requested scope.

Record model adapters, exact capability cards/version/status/provider surfaces, source tiers/dates/deprecations/conflicts, rights status, desired audio and generation audio route, preserve/change contracts, visual output mode, reference asset bindings, storyboard/clean-frame separation, optional model input graph, missing/generated/external asset status, forbidden direct inputs avoided, E0-E6 status, risk notes, QA status, and next_recommended_skill: generation-qa.