MekongMind supports multiple LLM providers with a unified interface. Bring your own API keys — no vendor lock-in.

​
Supported Providers

ProviderEnv VariableModel Example
OpenAIOPENAI_API_KEYgpt-4o, gpt-4o-mini
AnthropicANTHROPIC_API_KEYclaude-sonnet-4, claude-haiku-3.5
GoogleGOOGLE_API_KEYgemini-2.5-pro, gemini-2.5-flash
MLX (Local)—Qwen3.6-35B, Llama-3.2-3B

​
Configuration

mekong config set llm.provider openai
mekong config set llm.model gpt-4o

​
Routing

You can configure fallback providers: 
mekong config set llm.fallback anthropic
If the primary provider fails, requests automatically route to the fallback.

​
Local Models (MLX)

On Apple Silicon (M1+), run models locally: 
pip install mekong-cli[mlx]
mekong config set llm.provider mlx
mekong config set llm.model mlx-community/Qwen3.6-35B

​
Custom Endpoints

Any OpenAI-compatible endpoint works: 
mekong config set llm.provider openai
mekong config set llm.base_url https://my-custom-endpoint.com/v1
mekong config set llm.api_key sk-my-key