Skip to main content
Mirobody API is the managed way to use Mirobody. It’s the same open-source engine — we run it for you and wrap it in an OpenAI-compatible /v1 API, with API keys, usage metering, and a developer console. Point any OpenAI SDK at the base URL, pass an mb_live_* key, and the agent answers from each end-user’s real, structured health data, with traceable citations.
Built on the open source. This platform is a customized, managed layer on top of the open-source Mirobody engine — the same core, minus the infrastructure. You don’t build, host, or operate anything: get a key and call /v1. Everything the hosted agent does, the open-source engine does too — the API just makes it turnkey.

Two ways to use Mirobody

Pick the path that fits you — the engine underneath is the same.

Managed API — you are here

Get an mb_live_* key and call /v1/chat/completions. We host the engine, storage, and models. Best for shipping fast without running infrastructure.

Self-host the open source

Build the C++ engine and run it yourself — on a server or fully on-device. Full control; your data never has to leave your machine.

Why the managed API

  • OpenAI-compatible — reuse your existing OpenAI SDK; just change the base URL and key.
  • Answers from real data, cited — the agent reads each end-user’s health records and shows the sources it used.
  • Multi-tenant by design — one key, a fully isolated Subject per end-user.
  • No infrastructure — managed keys, storage, models, a console, and billing. Nothing to deploy.

Start here

Quickstart

Sign in, create a key, and make your first /v1 call in minutes.

API Reference

Auth, multi-tenancy, retention, and every /v1 endpoint.

Regions

Which cluster to use and what’s live today.

Compliance

PIPL / HIPAA / GDPR posture; request a DPA.