co/core
Developer docsDocs
Jump to
Inference

Inference API

OpenAI-compatible chat completions. Point your client at https://cocore.dev/api/v1, send a co/core API key as the Bearer token, and requests route to an attested provider on the network.

Base URLhttps://cocore.dev/api/v1

What you get

co/core exposes an OpenAI-compatible HTTP API at https://cocore.dev/api/v1. Use it with the OpenAI SDK, curl, or any client that speaks /v1/chat/completions. Each request is matched to an online provider and settled with a signed receipt.

Requests can carry text or images: vision-capable models accept inline base64 or remote image_url parts. See image input for the request shape and limits.

Before your first request

  1. Follow the quickstart to create an API key and wire up your client.
  2. Pick a model from the models directory or use stub while testing.

Using an editor or agent

OpenCode, Cursor, and Claude Code each need slightly different setup. See the tool setup pages for step-by-step configuration. Community-maintained extensions are listed on the community tools page.