OpenAI-compatible proxy with full observability

AI Gateway

Included in all plans

Route every LLM request through a single endpoint. Aethon handles provider routing, token counting, cost calculation, semantic caching, and request logging — with less than 50ms P99 overhead.

• — Drop-in replacement for OpenAI SDK
• — Multi-provider support (OpenAI, Anthropic, Google)
• — Per-request cost and latency tracking
• — API key authentication and rate limits
• — Dual-write to SQLite and ClickHouse