OpenCode with OpenCheese

← API quickstartView as Markdown

OpenCode connects to OpenCheese as an OpenAI-compatible provider at https://api.opencheese.dev/v1. OpenCode 1.x and 2.x read the same configuration.

Recommended: Auto-sync

With Auto-sync, OpenCode downloads its OpenCheese settings by itself: models you switch on or off on the Prices page, and new models, arrive without copying any file.

1. Sign in, open Configuration, choose the OpenCode tab and go to the Auto-sync steps.
2. Pick the API key OpenCode should use, enter that key's password and create your personal link.
3. Run the one command the page shows, once, in a terminal:

opencode auth login https://opencheese.dev/c/YOUR_PERSONAL_LINK

OpenCode stores the key and loads its configuration from OpenCheese from then on. The login runs a small Node.js command, so Node.js 18 or newer must be on your PATH (it is when you installed OpenCode with npm).

• Updates: OpenCode 2 picks up changes within 10 minutes, or at once with opencode reload (best while it is idle). OpenCode 1 picks them up when it starts, and can only start while OpenCheese is reachable.
• Your own settings: keep a small opencode.json with only your own things, such as an mcp section. Do not list models in it: models listed there stay visible even after you switch them off. Delete or trim an opencode.json you copied earlier.
• The link is a password: anyone who has it can use the key. It is shown once. Revoke it under Your personal links; an OpenCode that already logged in keeps the key, so disable or delete the API key to stop it. After changing the key's password, run the login again.

Or copy the generated file

Sign in, open Configuration, choose the OpenCode tab, pick an API key and confirm your password, then merge the generated file into your user configuration:

The generated file lists every text model your key can use, with its display name, context and output limits, per-token costs (OpenCode shows the cost of each message), reasoning variants and the right AI SDK package for each model. Prefer it to a hand-written one, and copy it again after your plan or model list changes.

Manual setup with one model

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "opencheese": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "OpenCheese",
      "options": {
        "baseURL": "https://api.opencheese.dev/v1",
        "apiKey": "{env:OPENCHEESE_API_KEY}"
      },
      "models": {
        "MODEL_ID": { "name": "Model name", "variants": {} }
      }
    }
  },
  "model": "opencheese/MODEL_ID"
}

• baseURL must end in /v1.
• {env:OPENCHEESE_API_KEY} reads the key from your environment; the generated file contains the key itself.
• Keep "variants": {} on every model you add by hand. Without it OpenCode 2.x offers low, medium and high reasoning variants of its own, which a model may not support.
• Take the model IDs from GET /v1/models with your key (see Models).

OpenCode 2

OpenCode 2 is the recommended version. Step 2 of the OpenCode tab switches between OpenCode 2 (the default) and OpenCode 1, which is legacy and supported until 1 November; both read the same settings. Moving up from OpenCode 1: OpenCode 2 is a different package and moves your session history over the first time it starts, so back that history up first. The tab shows the exact commands, including the way back to OpenCode 1.

Usage and costs

Every request appears under Transactions in the dashboard, grouped as one session per conversation, with its token usage and cost.

Troubleshooting

• "Invalid or disabled OpenCheese API key": check the key on the API keys page and that OPENCHEESE_API_KEY is set where OpenCode starts. With Auto-sync, run the login again.
• A model is missing: models depend on your plan and key restrictions. With Auto-sync, restart OpenCode 1 or run opencode reload in OpenCode 2; with a copied file, copy it again.
• A switched-off model still shows: a model listed in your local opencode.json stays visible; remove it there.
• "Not included in your current plan": choose another model, or keep Use balance for Credit Models on in the Overview page to pay for it from your balance.