Decentralized GPU inference on Robinhood Chain — read the docs

Quick start

These are the two primary paths in the Inference Pool system. The user path sends an inference request. The provider path lists GPU capacity and earns from completed jobs.

Use Inference Pool

Choose a model, send a request through the app or API, and receive streamed tokens from the selected worker.

Request
POST /v1/inference
model: pool-open-8b
stream: true

Provide compute

Run a worker client, benchmark your machine, list supported models, and accept routed jobs.

Worker CLI
inferencepool worker start
inferencepool worker benchmark
inferencepool worker list