Otari

Otari is an LLM gateway built on any-llm and made by Mozilla.ai. It gives teams a single place to send requests to different model providers, so applications don't depend on one vendor's API. It's aimed at teams that want control over how their AI traffic is routed, tracked, and paid for.

You can run the open source gateway yourself, or use the hosted platform at Otari.ai, which adds management features on top.

• Unified routing: send requests to multiple providers through one interface
• Observability: see what's being called, by whom, and how it performs
• Budgets: set spending limits so costs don't surprise you
• Team management: control who has access to which models and keys
• Hosted option: use Otari.ai if you'd rather not operate the gateway

The open source core is the main draw. You can inspect the code, keep traffic in your own infrastructure, and avoid being locked into a single provider or proxy service. Mozilla.ai's backing and the any-llm foundation mean provider support comes from a library built for exactly this job.

It sits in the same space as LiteLLM, Bifrost, and Portkey AI Gateway. If you're comparing hosted routing services such as OpenRouter, Otari offers a self-hosted path alongside its managed one. You can browse more options in the AI gateways category.