Workers AI
Workers AI is Cloudflare’s inference service: open models hosted on the same network the rest of a flow runs on, called through a binding instead of over the public internet. In SwiftTune it is the block that runs a prompt through a model and returns the completion.
What it means
Because inference is a binding rather than a third-party HTTP call, there is no key to rotate in the flow, and the round trip stays inside Cloudflare’s network. The catalogue spans text generation, embeddings, transcription and image models, each identified by a model name you select on the block.
How SwiftTune uses it
The Workers AI block declares its input and output with a schema, so a prompt built from another block’s output is validated before the call is made and the completion is validated before it moves on. Routing the same call through AI Gateway instead swaps the provider without changing the shape of the flow.