co/core
Developer docsDocs
Jump to
Integrations

Community tools

Extensions, plugins, and integrations built by the community to connect co/core with your favorite tools. Official setup guides for OpenCode, Cursor, and Claude Code live on the inference docs.

Listed tools

These projects are maintained independently. If you build something that uses co/core, open a pull request on graze-social/cocore to add it here.

pi-cocore

GitHub

A Pi extension that adds co/core as a model provider. Run open-source models on Apple Silicon via co/core — directly inside pi.

  • pi
  • Apple Silicon
  • extension
Install
pi install git:github.com/willnewby/pi-cocore
Setup
  1. Get your API key from cocore.dev.
  2. Start pi — on first run you'll be prompted for the key.
  3. Or run /cocore-setup at any time to configure or change your key.

cocore-local-gateway

GitHub

Exposes locally-running MLX models from a co/core agent as an OpenAI-compatible API endpoint over TCP — bridging the agent's Unix sockets to localhost and overlay networks (ZeroTier, Tailscale) without routing through the co/core network.

  • OpenAI API
  • MLX
  • gateway
  • macOS
Install
git clone https://github.com/tenorune/cocore-local-gateway
cd cocore-local-gateway
cp .env.example .env
./install.sh
Setup
  1. Configure .env with the port, bind interfaces, and socket directory.
  2. Run install.sh to load it as a LaunchAgent (auto-starts on macOS).
  3. Verify with: curl -s http://127.0.0.1:1234/v1/models | python3 -m json.tool
  4. Point your OpenAI-compatible clients (OpenCode, pi, OFF GRID) at the gateway endpoint.