restai

Why we included this project

RESTai is a self-hosted AI-as-a-Service platform that wraps a wide range of model capabilities in a single REST API, which makes sense if you want to offer AI to your own team or customers without handing data to a third-party vendor. Beyond plain chat completions it bundles RAG over uploaded documents, SQL-to-natural-language against a connected database, ReAct agents that can drive a headless browser through MCP tools, and image and audio generation, so one deployment can cover several common workloads at the same time. The admin web UI adds teams, role-based access control, OAuth and LDAP logins, and per-project rate limiting with token tracking, the operational plumbing a small team would otherwise have to assemble itself. If you already run models through Ollama or vLLM, or through a cloud provider, the same interface can route requests to any of those backends, and the Docker image or pip installer gets a first project running quickly.