LLM Providers

RsClaw supports 15 pre-configured LLM providers, plus any OpenAI-compatible API.

Pre-configured Providers

Provider	Base URL	Models
Qwen (Alibaba)	dashscope.aliyuncs.com	qwen-turbo, qwen-plus, qwen-max
DeepSeek	api.deepseek.com	deepseek-chat, deepseek-reasoner
Kimi (Moonshot)	api.moonshot.cn	moonshot-v1-8k/32k/128k
Zhipu (GLM)	open.bigmodel.cn	glm-4, glm-4v
MiniMax	api.minimax.chat	abab6.5
GateRouter	api.gaterouter.com	Multi-model routing
OpenRouter	openrouter.ai/api	100+ models
Anthropic	api.anthropic.com	Claude 3/4 family
OpenAI	api.openai.com	GPT-4o, o1, o3
Google Gemini	generativelanguage.googleapis.com	Gemini Pro/Ultra
xAI (Grok)	api.x.ai	Grok-2
Groq	api.groq.com	Llama, Mixtral
SiliconFlow	api.siliconflow.cn	Various open models
Ollama	localhost:11434	Local models
Custom	user-defined	Any OpenAI-compatible API

Configuration

{
  models: {
    providers: {
      qwen: {
        apiKey: "${DASHSCOPE_API_KEY}",
        baseUrl: "https://dashscope.aliyuncs.com/compatible-mode/v1",
      },
      openai: {
        apiKey: "${OPENAI_API_KEY}",
      },
      custom_provider: {
        apiKey: "sk-xxx",
        baseUrl: "https://your-api.example.com/v1",
      },
    },
  },
  agents: {
    defaults: {
      model: {
        primary: "qwen/qwen-turbo",
        fallback: "openai/gpt-4o-mini",
      },
    },
  },
}

Auto-Registration

Providers are auto-registered from environment variables:

Env Var	Provider
OPENAI_API_KEY	OpenAI
ANTHROPIC_API_KEY	Anthropic
DASHSCOPE_API_KEY	Qwen
DEEPSEEK_API_KEY	DeepSeek
GOOGLE_API_KEY	Gemini

Features

• Failover: automatic retry with exponential backoff
• Model fallback chains: primary -> fallback -> image fallback
• Thinking budget: configurable thinking level (low/medium/high)
• Token tracking: per-session and per-agent usage
• Vision auto-detect: auto-detect image-capable models