Skip to main content

​
Available Presets

Rusty ships with presets for popular LLM providers. Use --preset to select one:
PresetAPI BaseDefault Model
xiaomihttps://token-plan-cn.xiaomimimo.com/v1mimo-v2.5-pro
kimihttps://api.kimi.com/coding/v1/kimi-k2.6
openaihttps://api.openai.com/v1gpt-4o
deepseekhttps://api.deepseek.comdeepseek-v4-pro
ollamahttp://localhost:11434/v1llama3

​
Usage

# Use OpenAI with an API key
rusty --preset openai --api-key sk-...

# Use local Ollama
rusty --preset ollama

# Use DeepSeek
rusty --preset deepseek --api-key sk-...

​
Custom Endpoints

For providers not listed as presets, set --api-base and --model directly: 
rusty --api-base https://your-provider.com/v1 --api-key YOUR_KEY --model your-model
Any OpenAI-compatible API endpoint works, including:
  • Azure OpenAI
  • Together AI
  • Groq
  • Fireworks AI
  • Self-hosted vLLM or text-generation-inference

​
Presets vs Settings

When using --preset, the preset’s api_base and default_model override what is in ~/.rusty/settings.json. You can still override individual fields with CLI flags: 
# Use Xiaomi preset but with a different model
rusty --preset xiaomi --model mimo-v2-lite --api-key YOUR_KEY

​
Configuring a Default Provider

To set a default provider without using flags every time, edit ~/.rusty/settings.json: 
{
  "api_base": "https://api.openai.com/v1",
  "default_model": "gpt-4o",
  "credential_store": "keyring"
}
Then run with just: 
rusty