Overview
Lunaris AI API is a unified gateway for multiple AI providers. Instead of managing separate API keys, SDKs, and endpoints for each provider, Lunaris exposes a single OpenAI-compatible interface that routes requests intelligently across providers.
Key Features
- Unified Interface: One API key, multiple providers (Groq, Mistral, Nvidia, Gemini, OpenAI).
- Dynamic Model Fetching: Always up-to-date model lists queried directly from upstream providers.
- Automatic Failover: Zero downtime—if a provider fails, Lunaris switches instantly.
- Code Completion (FIM): Specialized endpoint for Fill-In-the-Middle IDE support.
- CLI Client: Built-in terminal client for rapid testing and chat.
Architecture
Built on the robust Tokio ecosystem with Axum as the HTTP framework. Each provider is a modular backend that implements a common trait, making it trivial to add new providers. The router layer handles provider selection (auto or manual), fallback chains, and request/response transformation.
API Endpoints
| Endpoint | Description |
| ------------------- | ------------------------------------------- |
| POST /chat | Chat completions with auto provider routing |
| POST /completions | Code FIM (Fill-In-the-Middle) completions |
| GET /models | List available models per provider |
Motivation
Managing multiple AI provider APIs is fragmented and brittle. Lunaris was built to provide a single, reliable entry point that abstracts provider differences, handles rate limits gracefully, and ensures uptime through intelligent fallback routing.
Last updated on July 8, 2026 at 8:07 AM UTC+7. See Changelog