Providers & Model Configuration 🧠

MalikClaw uses a protocol-first approach to model management. This allows you to integrate dozens of different LLM providers using a unified configuration format.

📋 The `model_list` System

All models are defined in the model_list array in your config.json. Each entry maps a friendly alias (model_name) to a specific protocol and provider.

Core Configuration Fields

| Field | Description | |-------|-------------| | model_name | A unique alias you use to refer to this model. | | model | The protocol/identifier (e.g., openai/gpt-4o). | | api_key | Your API credential for that provider. | | api_base | (Optional) The base URL for the API endpoint. | | auth_method | (Optional) Set to oauth or token for session-based auth. |

🌐 Supported Protocols

MalikClaw supports a wide range of protocols. If a provider isn't listed, it usually works via the openai/ protocol.

| Protocol Prefix | Target Provider | Typical API Base | |-----------------|-----------------|------------------| | openai/ | OpenAI | https://api.openai.com/v1 | | anthropic/ | Anthropic (SDK) | https://api.anthropic.com/v1 | | anthropic-messages/ | Anthropic (Native) | https://api.anthropic.com | | gemini/ | Google Gemini | https://generativelanguage.googleapis.com/v1beta | | antigravity/ | Google Cloud Code Assist | (Managed internally) | | groq/ | Groq | https://api.groq.com/openai/v1 | | deepseek/ | DeepSeek | https://api.deepseek.com/v1 | | openrouter/ | OpenRouter | https://openrouter.ai/api/v1 | | ollama/ | Ollama (Local) | http://localhost:11434/v1 | | zhipu/ | Zhipu AI (GLM) | https://open.bigmodel.cn/api/paas/v4 | | modelscope/ | ModelScope | https://api-inference.modelscope.cn/v1 |

🦅 Antigravity (Google Cloud Code Assist)

Antigravity provides access to high-end models (including Claude 3.5 Opus) via Google's infrastructure. It requires a specific OAuth setup.

{
  "model_name": "gemini-ultra",
  "model": "antigravity/gemini-2.0-flash",
  "auth_method": "oauth"
}

Setup: Run malikclaw auth login --provider google-antigravity to authenticate.

⚡ High Availability & Fallbacks

MalikClaw can automatically switch to backup models if your primary provider is down or rate-limited.

{
  "agents": {
    "defaults": {
      "model": "primary-gpt-4o",
      "fallbacks": ["secondary-claude-sonnet", "backup-deepseek"]
    }
  }
}

🤖 GitHub Copilot

MalikClaw can hook into an existing GitHub Copilot subscription.

Connection Modes:

gRPC (Remote): Connects to a running Copilot CLI server.
stdio (Local): Spawns the copilot binary directly.

{
  "model_name": "my-copilot",
  "model": "github-copilot/gpt-4o",
  "connect_mode": "stdio",
  "api_base": "copilot"
}

🛠️ Protocol-Specific Options

thinking_level: For Claude models, set to low, medium, or high.
max_tokens_field: Specify custom field names for token limits (e.g., max_completion_tokens).
request_timeout: Integer seconds before timing out an LLM request.