# Otari

> Apache-2.0 self-hosted OpenAI-compatible LLM gateway

**Category:** AI Agents  
**Pricing:** open-source  
**URL:** https://otari.ai/docs/gateway  
**GitHub:** https://github.com/mozilla-ai/otari  
**Added:** 2026-10-19

---

## The Hook

Otari is an LLM gateway from Mozilla.ai that sits between your apps and model providers: one OpenAI- and Anthropic-compatible endpoint in front of 40+ providers, with revocable virtual keys and budgets checked before a request is sent.

## What's Good

Apache-2.0 (mozilla-ai/otari), release v0.28.0 (2026-10-16). Provider credentials stay behind the gateway, API keys can be scoped by user, workspace and model, and usage is recorded after each call. Local routing policies handle failover and weighting, and optional add-ons cover code execution, web search, MCP and guardrails. Provider calls go through Mozilla.ai's any-llm library, any OpenAI client works by changing base_url, and otari hook / otari import claude-code wire it into a coding agent. ★~670.

## The Catch

It is younger than LiteLLM and still pre-1.0, so expect breaking changes. The one-line Docker quick start keeps SQLite inside the container and loses everything when it stops; a persistent setup means Compose with PostgreSQL. The README also offers a hosted control plane at otari.ai, so check which features you need run fully standalone.

## Verdict

A clean, self-hosted way to put keys and budgets in front of many LLM providers — pin a version while it is pre-1.0.