Repolode
Submit
apocas

apocas/restai

Open-source AIaaS platform for building RAG, agent, and visual-logic AI projects and exposing them through a REST API with a built-in web UI.

View on GitHubWebsite
Stars
517
Forks
113
Open issues
39
Last commit
6 days ago

What is it?

RESTai is a self-hosted AI-as-a-Service platform. You create AI projects of different types (RAG, agent, block, or inference), each with its own LLM, system prompt, tools and settings, and then consume them through a REST API or the built-in React dashboard. It supports a wide range of model providers, including OpenAI, Anthropic, Ollama, Gemini, Azure, AWS Bedrock and vLLM, plus OpenAI-compatible endpoints.

The platform covers the operational side of running AI features for teams. It includes token, cost and latency analytics, prompt versioning with restore, evaluation runs using DeepEval metrics, guardrail projects for input and output checks, per-project rate limiting, an audit log, team-level permissions and white-label branding. Authentication includes OAuth/LDAP, TOTP two-factor login and API keys.

RAG projects ingest documents and can sync knowledge from web URLs, S3, Confluence, SharePoint and Google Drive. Other capabilities include an MCP server for clients such as Claude Desktop, a Telegram, Slack and WhatsApp integration, an embeddable chat widget, a WordPress plugin, and image, audio and classification tools. It can be installed from PyPI, run with Docker, or deployed to Kubernetes with a Helm chart.

Key features

  • 01Create RAG, agent, block and inference projects, each with its own LLM and configuration
  • 02Connects to many LLM providers, including local models through Ollama and vLLM
  • 03Built-in analytics for token usage, cost and latency, with a React dashboard
  • 04Prompt versioning with history and one-click restore, plus evaluation runs
  • 05MCP server support and a Blockly-based visual logic builder
  • 06OpenAI-compatible endpoints for chat, image generation and audio transcription

Use cases

  • Hosting an internal question-answering bot over company documents with scheduled sync from S3 or Confluence
  • Giving a team shared, access-controlled access to multiple LLM providers with usage and cost tracking
  • Embedding a support chat bot on a website or connecting one to Telegram, Slack or WhatsApp
  • Running local open-source models behind an OpenAI-compatible API for existing SDK-based code

Quick start

  1. 1

    Install from PyPI

    pip install restai-core
  2. 2

    Create database and admin user

    restai init
  3. 3

    Run database migrations

    restai migrate
  4. 4

    Start the server, then open http://localhost:9000/admin (admin / admin)

    restai serve

Things to know

  • Default admin credentials are admin / admin and must be changed.
  • The feature set is large and heavy, so self-hosting needs real setup effort.

In collections

Private AI knowledge bases and RAGOpen-source AI agent frameworks & appsSelf-hosted appsRun LLMs locally

Topics

#llm#rag#ai-agent#openai-compatible#ollama#fastapi#self-hosted#embeddings#blocky#langchain

FAQ

What is restai?
Open-source AIaaS platform for building RAG, agent, and visual-logic AI projects and exposing them through a REST API with a built-in web UI. RESTai is a self-hosted AI-as-a-Service platform. You create AI projects of different types (RAG, agent, block, or inference), each with its own LLM, system prompt, tools and settings, and then consume them through a REST API or the built-in React dashboard. It supports a wide range of model providers, including OpenAI, Anthropic, Ollama, Gemini, Azure, AWS Bedrock and vLLM, plus OpenAI-compatible endpoints.
Is restai free to use? What license does it have?
restai is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is restai still maintained?
Actively maintained. The last commit to restai was on Oct 30, 2026.
What can I use restai for?
Hosting an internal question-answering bot over company documents with scheduled sync from S3 or Confluence Giving a team shared, access-controlled access to multiple LLM providers with usage and cost tracking Embedding a support chat bot on a website or connecting one to Telegram, Slack or WhatsApp Running local open-source models behind an OpenAI-compatible API for existing SDK-based code
How do I install restai?
Install from PyPI: pip install restai-core Create database and admin user: restai init Run database migrations: restai migrate Start the server, then open http://localhost:9000/admin (admin / admin): restai serve
Where can I get restai?
The source code and installation instructions are on GitHub: https://github.com/apocas/restai

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects