Meridian Documentation
Meridian is an open-source LLM API gateway and proxy. Add your provider API keys once, generate proxy keys, and call any model through a unified OpenAI-compatible API. Meridian handles routing, encryption, logging, analytics, and A/B experiments behind the scenes.
Who is Meridian for?
- Developers who use multiple LLM providers and want a single API interface.
- Teams that need visibility into LLM usage, cost, and latency across projects.
- Security-conscious organizations that want API keys encrypted at rest and never exposed in app code.
- Experimenters who want to A/B test models without changing application logic.
Key Features
Proxy Keys
Generate revocable proxy keys so your real API keys never touch application code.
Unified API
One OpenAI-compatible endpoint for OpenAI, Anthropic, and Google Gemini models.
A/B Experiments
Route traffic between models with configurable splits to compare cost and quality.
Usage Analytics
Per-request logging with token counts, latency, cost estimates, and a real-time dashboard.
AES-256 Encryption
Provider keys encrypted at rest with AES-256-GCM and decrypted only during request execution.
Self-Hostable
Deploy to Vercel in minutes or run on your own infrastructure. Apache 2.0 licensed.
Get Started
New to Meridian? The quickstart guide walks you through signup, adding a provider key, creating your first proxy key, making an API call, and viewing usage — all in under five minutes.
Quickstart Guide →API Reference
Meridian exposes an OpenAI-compatible REST API. All endpoints accept a Meridian proxy key via the Authorization: Bearer header.
- POST /api/v1/chat/completions — Chat and text generation
- POST /api/v1/images/generations — Image generation
- POST /api/v1/embeddings — Text embeddings
- POST /api/v1/audio/speech — Text-to-speech
- GET /api/usage — Usage analytics and logs