Creativeclaw Generate Video

Generate Video

Read shared execution guidance once per task before using tools. It covers existing authorization, model discovery, optional cost checks, imports, and recovery.

Create one controlled video clip from text, a start frame, an optional end frame, or other model-supported references. Use the planning, UGC, or film skills when the deliverable is a larger production. Use creativeclaw-render-html-video only when the user explicitly requests HTML/HyperFrames/code-driven rendering; use creativeclaw-add-video-intro-outro for video bookends.

Workflow

1. Define the clip's purpose, aspect ratio, duration, subject, one primary action, camera move, visual continuity, dialogue or sound, and required end state.
2. Search existing assets before importing new references. A still is a start frame only when it should define the opening composition.
3. When the user asks for examples, references, styles, or similar concepts—or an open brief would materially benefit from choosing among concrete directions—use creativeclaw-find-examples. Filter by output_type: "video", then load only the selected result. Do not search before every clip.
4. Use list_models({ category: "video" }) when choosing a model; for a known selection, use get_model_params directly. Reuse its current-task durations, resolutions, operations, and reference contract.
5. Use estimate_generation with operation: "video" only when the user asks about cost, balance, affordability, or sets a budget. Treat returned alternatives as options; preserve explicitly chosen models, durations, and quality. Estimate-only requests do not authorize generation.
6. State consequential settings briefly and proceed within the requested scope. Do not ask again when the user already requested the generation or approved that production stage.
7. Write one chronological prompt: opening frame, subject action, camera behavior, environmental motion, audio or dialogue, ending frame, and exclusions.
8. Call generate_video; use check_job only when another tool needs the completed URL or no inline viewer is monitoring the job.
9. Inspect identity, anatomy, product fidelity, timing, camera motion, dialogue sync, and ending continuity. Revise one variable at a time.

Model routing

• Default to video/gemini-omni-flash for the best general balance of speed, quality, native audio, and reference-aware generation.
• Use video/seedance-2.5 for premium cinematic or reference-rich clips and longer shots where supported.
• Use video/seedance-2.0-mini, presented to users as Seedance Mini, for economical drafts.
• Use video/minimax-h3-max for fast cinematic motion and native-audio work.
• Use video/minimax-h3-max-turbo, presented as H3 Max Fast, when speed and iteration cost matter most.
• Honor an explicit model request, and load that model's specialist skill for exact prompt and reference syntax.

Reference rules

• image_url is the start frame. last_frame_url is the desired end frame when the selected model exposes it.
• image_urls, video_urls, and audio_urls are model-specific reference arrays. There is no universal three-reference requirement.
• A saved character_id supplies the Character image as the start frame only when no explicit image_url is provided. If a storyboard is the start frame and identity must also be referenced, add the Character image through a supported reference field.
• Preserve literal reference tokens such as Seedance @Image1, exact dialogue, or timecodes by setting agentic_prompting: false.
• Use operation only for the exposed modes: retake, extend, reframe, audio_to_video, or animate_character.

Prompt shape

Prefer one subject action and one camera idea per clip. Describe what happens over time, not a pile of adjectives. Include exact spoken words only when needed, and specify what must not change. For multi-shot continuity, first create a storyboard and clean reference frames with creativeclaw-plan-video.

Conduct the workflow in the user's language and preserve quoted dialogue exactly. Confirm the chosen model supports the requested spoken language before relying on native audio.