Back to Gap Map Product / UX / UI & API

Otari

Mozilla AI
open source / Overall score: 1.6

Open, self-hosted OpenAI-compatible LLM gateway from Mozilla AI. It puts one endpoint in front of many providers with virtual API keys, budget enforcement and usage tracking, and ships server-side model-agnostic tools including sandboxed code execution in a Docker-isolated Python REPL. Built on Mozilla AI's any-llm, with a hosted platform running on the same core.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI, core)
source
public
self-host
yes(Docker Compose)
service
hosted Otari.ai platform on the same core

Otari's gateway is Apache-2.0 and self-hostable via Docker Compose, and Mozilla AI's hosted Otari.ai platform runs on that same published code rather than on a separately licensed piece. A hosted convenience alongside a complete open core is not the same as gating: nothing in the source is withheld or paywalled.

Adoption

1 low confidence
1.0

Otari's repository is very new and its star count remains low, so the adoption signal here is thin; no download, deployment or customer figure has appeared yet. Mozilla AI's any-llm library, which Otari is built on, has far more traction, but it is a separate product.

Capability

3 medium confidence
3.0

Otari is an OpenAI-compatible gateway to many providers behind one API, with virtual keys, per-user budgets, usage tracking and sandboxed server-side code execution.

  • https://blog.mozilla.ai/otari-own-your-ai-stack/ recorded 2026-09-29

    Launch post feature lists: "Virtual API keys: Hashed, named, optionally-expiring keys bound to a user", "Usage and spend tracking: Real-time cost calculation across providers", "Sandboxed code execution. A Docker-isolated Python REPL, invoked server-side when a model needs to run code", and "Per-user spending caps with configurable reset windows".

  • https://github.com/mozilla-ai/otari recorded 2026-08-13

    40+ providers, virtual keys, budgets, usage tracking, sandboxed code execution

Verified 2026-09-27