RESTai is an open-source AI-as-a-Service platform hosted on GitHub, with about 520 stars under the Apache-2.0 license and active updates in 2026. It lets you create AI projects once and then consume them through a simple REST API, instead of wiring model calls into every application separately.
The platform supports many public LLMs as well as local models served by Ollama, vLLM, and similar runtimes. It handles precise embeddings usage with tuning and analytics, prompt versioning, and live chat deployment. Beyond text, RESTai includes built-in image and audio generation with dynamically loaded generators, plus a block-based graphical language for composing logic visually. It is built on FastAPI with LangChain and LlamaIndex, installs from PyPI as restai-core with a pre-built React admin frontend, and ships Docker- and Kubernetes-ready. A public demo runs on the project’s cloud site.
Getting started is quick: the PyPI package bundles the pre-built React frontend, so no Node.js build step is needed, and a short init, migrate, and serve sequence brings up the admin dashboard. A restricted public demo account lets you browse and chat with existing projects before installing anything, and the same codebase scales from a single local process to Docker or Kubernetes deployments using an environment file for configuration.
RESTai is free and open source under the Apache-2.0 license for self-hosting. A hosted cloud version is available through restai.cloud for teams that prefer not to run the platform themselves.
It suits developers and small teams who want one self-hosted gateway for LLM, embedding, image, and audio features across several apps, with prompt management and analytics included rather than rebuilt per project.