Files

1.2 KiB

Workers AI API

Fetch the selected model's page from the model catalog for request fields, output format, dimensions, and examples. Text, embeddings, images, audio, and translation do not share one response schema.

Task Documentation
Invoke inference through a Worker binding Workers bindings
Invoke inference over HTTP REST API reference
Stream text or use SDK abstractions Vercel AI SDK, or the selected model's streaming example
Define tools and handle tool results Function calling
Request structured output JSON mode

Use the response and stream format documented for the chosen integration. Do not assume native binding streams are parsed objects or apply OpenAI response parsing to every model. Check the model's batching support and limits before combining inputs in one request.