Skip to content

docs: add Wan 3.0 guide - #530

Draft
tombystrican wants to merge 3 commits into
mainfrom
tomb/wan-3-0-guide
Draft

tombystrican wants to merge 3 commits into
mainfrom
tomb/wan-3-0-guide

Conversation

@tombystrican

@tombystrican tombystrican commented Sep 10, 2026

Copy link
Copy Markdown
Contributor

Summary

Adds guides/media/wan-3-0.mdx, a Venice-side guide for the Wan 3.0 family, and links it from the Video guides group in docs.json and from llms.txt.

Cross-checked against the video model definitions, the media limits, the public /video/queue schema, and the video pricing logic on outerface main.

Covers:

  • the 12 public model IDs (standard / Prime / Pro / Prime Pro × T2V / I2V / R2V) and which tier takes document references
  • Omni Reference vs First & Last Frames (mutually exclusive input modes, separate variants on Venice)
  • request parameters: 20k-char prompt, 2–30 s ladder, 480p–4K, adaptive ratio, audio, and what is not supported (negative_prompt, prompt rewriting, smart duration)
  • @Image1 / @Video1 / @Audio1 reference syntax and ordinal binding
  • workflows with prompt patterns: subject reference, video temporal reference, audio reference, document/webpage, video edit (9 edit types), video extend (forward / backward / both)
  • multimodal limits incl. the 15 s aggregate reference-video cap and the 30 s input+output cap
  • pricing via /video/quote with reference_video_total_duration
  • 8 complete curl examples, polling, troubleshooting

No provider or vendor names appear in the guide. Staff-only Wan 3.0 Enhanced is intentionally not mentioned.

Notes for review

  • No hard per-second prices are stated; the guide defers to /video/quote like the Seedance guide does.
  • Example prompts are original or shortened adaptations; no third-party media is embedded.

🤖 Generated with Claude Code

@mintlify

mintlify Bot commented Sep 10, 2026

Copy link
Copy Markdown
Contributor

Preview deployment for your docs. Learn more about Mintlify Previews.

Project Status Preview Updated
veniceai 🟢 Ready View Preview Sep 10, 2026, 2:18 AM

Venice-side guide for the Wan 3.0 family (standard, Prime, Pro, Prime Pro):
variants, Omni Reference vs First & Last Frames, request parameters, the
@ImageN/@VideoN/@audion reference syntax, reference/edit/extend/document
workflows with prompt patterns, multimodal limits, pricing via /video/quote,
complete curl examples, and troubleshooting. Cross-checked against the video
model definitions and the public queue schema.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
- aspect_ratio and duration are required; resolution defaults to 1080p
- the 30 s input+output cap is enforced at dispatch: queue accepts the job,
  retrieve returns 400 and the pre-charge is refunded
- add the 'image file is corrupted or unreadable' troubleshooting entry
- drop remaining vendor/upstream wording

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Outerface now stores the provider's message on the task instead of the generic 'Invalid request parameters', so the troubleshooting entry quotes that message and shows the response body.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant