Goose integration

Register Modellane as a custom provider in the Goose agent with a small JSON file, then choose a model for your sessions.

Goose supports custom OpenAI-compatible providers. Describe Modellane in one JSON file, or add it from the provider settings, and pick a model.

Goose is an open-source agent from Block that runs on your machine, as a desktop app or a CLI. It supports custom providers with an OpenAI engine, which call the Chat Completions API (POST /v1/chat/completions).

Add the provider#

In the desktop app, open Settings, then Models, Configure providers and Add Custom Provider. In the CLI, run goose configure and choose Custom Providers. Both write a JSON file; you can also create it yourself at ~/.config/goose/custom_providers/modellane.json (on Windows, %APPDATA%\Block\goose\config\custom_providers\):

JSON

{
  "name": "modellane",
  "engine": "openai",
  "display_name": "Modellane",
  "api_key_env": "MODELLANE_API_KEY",
  "base_url": "https://usemodellane.com/v1/chat/completions",
  "models": [
    { "name": "lane-1", "context_limit": 128000 }
  ],
  "supports_streaming": true,
  "requires_auth": true
}

The base_url holds the full chat completions address, as in the Goose documentation. Replace context_limit with the context length of the model from Models & pricing, and add one entry per model you want to use.

Export the key, then pick Modellane and the model in Goose:

Shell

export MODELLANE_API_KEY="your-api-key"
goose session

Verify#

Start a session and send "Reply with the word ready." A reply means the provider works. An authentication error points at the key, and model_not_found at the model id.

Images#