Configuration

Configure Browser, Ollama, and Rapid MLX in OpenGPT0.

OpenGPT0 keeps configuration focused on the built-in Browser, Ollama, and Rapid MLX providers.

Config Files

OpenGPT0 reads configuration from global and project config files. Project config overrides global config.

{
  "$schema": "https://opengpt0.com/config.json",
  "model": "browser/chatgpt"
}

Browser

Browser is the product default and does not need an API key in config.

{
  "$schema": "https://opengpt0.com/config.json",
  "model": "browser/chatgpt"
}

Browser variants can be selected from the model picker: instant, medium, high, extra-high, and pro.

Ollama

Ollama uses the OpenAI-compatible local endpoint and ships with qwen3.5:9b-mlx as the built-in local model.

{
  "$schema": "https://opengpt0.com/config.json",
  "provider": {
    "ollama": {
      "options": {
        "baseURL": "http://127.0.0.1:11434/v1",
        "apiKey": "ollama"
      }
    }
  },
  "model": "ollama/qwen3.5:9b-mlx"
}

To add another local Ollama model, pull it first and add the model under the ollama provider.

ollama pull qwen3.5:9b-mlx

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

Rapid MLX uses a local OpenAI-compatible endpoint at http://127.0.0.1:8080/v1. OpenGPT0 treats it as a separate built-in local provider and expects the local service to expose /health, /v1/models, and /v1/chat/completions.

{
  "$schema": "https://opengpt0.com/config.json",
  "provider": {
    "rapid-mlx": {
      "options": {
        "baseURL": "http://127.0.0.1:8080/v1",
        "apiKey": "rapid-mlx"
      }
    }
  },
  "model": "rapid-mlx/qwen3.5-9b-4bit"
}

Provider Filters

Provider filters should only reference the built-in providers.

{
  "enabled_providers": ["browser", "ollama", "rapid-mlx"]
}