Online documentation:
| Provider |
Type |
Tier |
Best For |
| Ollama |
ollama |
Free |
Local, private, no API costs |
| LM Studio |
lmstudio |
Free |
Local with GUI model management |
| Anthropic |
anthropic |
Pro |
Claude models — best coding quality |
| OpenAI |
openai |
Pro |
GPT models — versatile, fast |
| DeepSeek |
deepseek |
Pro |
Affordable, strong reasoning |
| Groq |
groq |
Pro |
Ultra-fast inference |
| OpenRouter |
openrouter |
Pro |
Access to 100+ models via one API |
leanctl init:
leanctl — Setup Wizard
This will configure your AI provider in ~30 seconds.
┌─────────────────────────────────────────────────────┐
│ Choose your AI provider: │
│ │
│ 1. Anthropic Claude models (recommended) │
│ → needs API key from anthropic.com │
│ │
│ 2. OpenAI GPT models │
│ → needs API key from openai.com │
│ │
│ 3. Ollama Local models (free, private) │
│ → needs Ollama installed locally │
│ │
│ 4. DeepSeek DeepSeek models (affordable) │
│ → needs API key from deepseek.com │
│ │
│ 5. Groq Ultra-fast inference │
│ → needs API key from groq.com │
│ │
│ 6. Custom Any OpenAI-compatible API │
│ → provide your own URL + key │
└─────────────────────────────────────────────────────┘
Choice [1-6] (1):
missing lmstudio, openrouter, and using custom with own base URL + optional key gives error in runtime:
openai_compatible not supported.
Online documentation:
leanctl init:missing lmstudio, openrouter, and using custom with own base URL + optional key gives error in runtime:
openai_compatible not supported.