Skip to content

docs: add WaveSpeed AI provider page - #962

Open
chengzeyi wants to merge 2 commits into
BerriAI:mainfrom
chengzeyi:docs/wavespeed-provider
Open

docs: add WaveSpeed AI provider page#962
chengzeyi wants to merge 2 commits into
BerriAI:mainfrom
chengzeyi:docs/wavespeed-provider

Conversation

@chengzeyi

Copy link
Copy Markdown

Adds a provider page for WaveSpeed AI covering image generation, video generation and the OpenAI-compatible chat endpoint, with SDK and proxy examples for each surface plus the supported parameter tables. It also explains how the asynchronous prediction API maps onto each one: LiteLLM polls internally for images, while video generation returns a task id and you drive video_status and video_content yourself.

WAVESPEED_API_KEY and WAVESPEED_API_BASE are added to the environment variables reference in docs/proxy/config_settings.md, and the page is registered in sidebars.js after the WatsonX category.

This is the docs half of BerriAI/litellm#37627, which adds the provider itself. That PR is currently red on code-quality and documentation, both because tests/documentation_tests/test_env_keys.py reports WAVESPEED_API_BASE and WAVESPEED_API_KEY as read under ./litellm but documented nowhere; merging this fixes both. It also makes the https://docs.litellm.ai/docs/providers/wavespeed link that the code PR registers in provider_endpoints_support.json resolve instead of 404.

Verified by checking out this branch into docs/my-website of the code PR's branch and running the four documentation tests the CI jobs run, all of which pass:

test_env_keys exit=0
test_router_settings exit=0
test_api_docs exit=0
test_circular_imports exit=0

npm run build also succeeds and emits build/docs/providers/wavespeed/index.html. The build reports broken links, but every one of them is pre-existing on other pages; none are on or point to the new page.

Covers image generation, video generation and the OpenAI-compatible chat
endpoint, with SDK and proxy examples for each, plus the supported parameter
tables and how the asynchronous prediction API maps onto each surface.

Adds WAVESPEED_API_KEY and WAVESPEED_API_BASE to the environment variables
reference, which unblocks tests/documentation_tests/test_env_keys.py in the
main repo, and registers the page in sidebars.js.
@vercel

vercel Bot commented Aug 20, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
litellm Ready Ready Preview Aug 20, 2026 12:35pm

Request Review

Chat and media sit on different hosts, and the provider now ignores the chat
base on the media path rather than building an unreachable prediction URL, so
sharing the variable between the two surfaces is safe.
@chengzeyi

Copy link
Copy Markdown
Author

Cross-linking: BerriAI/litellm#37627 is blocked on this PR. Its code-quality and documentation jobs check out this repo's default branch into docs/my-website (no ref), then fail with Environment variables read under ./litellm but mentioned nowhere in the docs: ['WAVESPEED_API_BASE', 'WAVESPEED_API_KEY']. The rows added here are what satisfy that check — every other check on #37627 is green.

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