Providers

Limitcode supports these providers:

• OpenAI
• DeepSeek
• Anthropic
• Gemini
• OpenRouter
• Moonshot AI (Kimi)
• Ollama
• LM Studio

Cloud providers require an API key. Ollama and LM Studio run locally without a cloud API key. Model lists are fetched from each provider’s models endpoint and use a cached fallback when needed.

To configure a provider, run Limitcode: Setup Provider API Key from the Command Palette, then select a provider and model in the Limitcode chat.

Local models

For local inference, start Ollama or LM Studio before opening a chat. Configure the provider’s base URL when the default endpoint does not match your local setup.

Provider-specific settings

Use provider_base_urls for custom or self-hosted OpenAI-compatible endpoints:

{
    "provider_base_urls": {
        "openai_compatible": "http://localhost:1234/v1"
    }
}