Providers

OpenGPT0 intentionally keeps the provider surface small. The built-in product supports three provider paths:

• Browser: uses your ChatGPT web session through the dedicated Chromium profile. Select browser/chatgpt.
• Ollama: runs local models through Ollama’s OpenAI-compatible endpoint. Select an ollama/... model.
• Rapid MLX: runs local MLX models through a separate OpenAI-compatible endpoint. Select a rapid-mlx/... model.

Browser

Browser is the default OpenGPT0 path. The ChatGPT browser provider drives the ChatGPT webpage, while local file reads, shell commands, edits, and Git operations still go through OpenGPT0’s permissioned tools.

Use this model ID:

browser/chatgpt

Available Browser variants are instant, medium, high, extra-high, and pro.

Ollama

Start Ollama locally and pull the model you want to use:

ollama pull qwen3.5:9b-mlx

OpenGPT0 talks to Ollama through this default endpoint:

http://127.0.0.1:11434/v1

Optional config:

opengpt0.json

{
  "$schema": "https://opengpt0.com/config.json",
  "provider": {
    "ollama": {
      "options": {
        "baseURL": "http://127.0.0.1:11434/v1",
        "apiKey": "ollama"
      },
      "models": {
        "qwen3.5:9b-mlx": {
          "name": "Qwen3.5 9B MLX",
          "limit": {
            "context": 32768,
            "output": 8192
          }
        }
      }
    }
  },
  "model": "ollama/qwen3.5:9b-mlx"
}

Rapid MLX

Rapid MLX is exposed as a separate built-in local provider. OpenGPT0 expects the local runtime to expose an OpenAI-compatible endpoint at:

http://127.0.0.1:8080/v1

Built-in model:

rapid-mlx/qwen3.5-9b-4bit

Only Browser, Ollama, and Rapid MLX are built into OpenGPT0. Other provider integrations are intentionally not part of the default product.