Name of app / service
WaveSpeed AI (wavespeed) — https://wavespeed.ai — a hosted inference API for image and video generation. Models are addressed by id (e.g. bytedance/seedream-v5.0-pro, bytedance/seedance-2.5/text-to-video) through a uniform submit/poll REST contract.
Link to developer documentation
Is lack of support preventing you from moving forward, or do you have a workaround?
There is a workaround through the generic HTTP Request action, but generation is asynchronous: POST /api/v3/{model} returns a prediction id and the caller must poll GET /api/v3/predictions/{id}/result until a terminal status. Doing that by hand means users re-implement the polling loop in every workflow, and a mis-built retry on the submit call bills a duplicate paid task. A native app makes the key connection reusable and puts the submit/poll handling inside the action.
I'm on the WaveSpeed team and have the components built locally (wavespeed.app.mjs + 4 actions). I'm requesting the registry app entry plus custom API-key auth so I can open the components PR.
Auth is a single API key sent as Authorization: Bearer <key>. Please register the custom-auth field as api_key (the components read this.$auth.api_key).
Are there specific actions or triggers you'd like to see for this app? Please let us know here or use the Action and Trigger issue templates to open requests for each!
Actions:
- Generate Image — submit an image model and return the resulting URLs (default
bytedance/seedream-v5.0-pro).
- Generate Video — same for video models (default
bytedance/seedance-2.5/text-to-video), with the option to return the prediction id immediately rather than block a step on a long render.
- Run Model — generic escape hatch: any model id plus a free-form inputs object.
- Get Prediction — fetch a prediction by id, so a job that outlives a workflow step can be picked up by a later run.
The model dropdown is populated from the public catalog endpoint GET /api/v3/models, with free-text entry allowed so newly released models work without a component release.
No trigger is needed for the initial contribution.
Disclosure: I work at WaveSpeed and will contribute and maintain these components.
Name of app / service
WaveSpeed AI (
wavespeed) — https://wavespeed.ai — a hosted inference API for image and video generation. Models are addressed by id (e.g.bytedance/seedream-v5.0-pro,bytedance/seedance-2.5/text-to-video) through a uniform submit/poll REST contract.Link to developer documentation
https://api.wavespeed.aiIs lack of support preventing you from moving forward, or do you have a workaround?
There is a workaround through the generic HTTP Request action, but generation is asynchronous:
POST /api/v3/{model}returns a prediction id and the caller must pollGET /api/v3/predictions/{id}/resultuntil a terminal status. Doing that by hand means users re-implement the polling loop in every workflow, and a mis-built retry on the submit call bills a duplicate paid task. A native app makes the key connection reusable and puts the submit/poll handling inside the action.I'm on the WaveSpeed team and have the components built locally (
wavespeed.app.mjs+ 4 actions). I'm requesting the registry app entry plus custom API-key auth so I can open the components PR.Auth is a single API key sent as
Authorization: Bearer <key>. Please register the custom-auth field asapi_key(the components readthis.$auth.api_key).Are there specific actions or triggers you'd like to see for this app? Please let us know here or use the Action and Trigger issue templates to open requests for each!
Actions:
bytedance/seedream-v5.0-pro).bytedance/seedance-2.5/text-to-video), with the option to return the prediction id immediately rather than block a step on a long render.The model dropdown is populated from the public catalog endpoint
GET /api/v3/models, with free-text entry allowed so newly released models work without a component release.No trigger is needed for the initial contribution.
Disclosure: I work at WaveSpeed and will contribute and maintain these components.