by Cloudflare
Cloudflare Workers AI runs open-source machine learning models on serverless GPUs across Cloudflare's network. You call the models from Workers, Pages or the Cloudflare API without managing infrastructure.
APIMCP serverUsage-based
Step 1: Browse the model catalog and pick a model for your task.
Step 2: Add an AI binding to a Cloudflare Worker, or get an account ID and API token to use the REST API.
Step 3: Run the model from your Worker, Pages function or an HTTP call to the Cloudflare API.
Step 4: For large workloads, send a batch request with queueRequest set to true and keep the returned request ID.
Step 5: Poll with the request ID until the batch returns its responses.
Step 6: Optionally add AI Gateway or Vectorize for observability and retrieval.
You send a collection of inference requests in one call and the system queues them, returning a status of queued and a request ID. You poll with that ID to get the responses once processing finishes.
It suits large jobs with no human waiting, such as summarization or embeddings. Batched requests are guaranteed to be fulfilled eventually rather than failing when capacity is short.
Yes, through function calling. You define tools and the model returns JSON for the call, or you can use embedded function calling with the ai-utils package to run the function code alongside inference.
From Cloudflare Workers through an AI binding, from Pages, or from anywhere using the Cloudflare REST API.
Official Cloudflare platform skill covering Workers, Workers AI, storage, Agents SDK, networking, security, and deployment workflows.
InworldInworld provides realtime text-to-speech, speech-to-text, speech-to-speech and LLM inference APIs aimed at consumer apps such as companions, tutors and games.
Fireworks AIAI inference platform for running leading open models with serverless APIs, fine-tuning and dedicated deployments.
GensynGensyn builds AI that forecasts and verifiably checks its predictions against reality. Its developer tools include REE for reproducible inference, AXL for agent communication and SDKs for Delphi information markets.