中文

RESTai: AI projects consumed via a REST API

Create AI projects and consume them via a REST API, with a React dashboard that includes analytics.

Verified

Core facts

Read by a script from the official API, not rewritten by a model.

License
Apache-2.0
Languages
Python
Maintainer
apocas
Verified

What it is

RESTai is AIaaS (AI as a Service): create AI projects and consume them via a simple REST API. It includes a React dashboard with analytics, not just an API. Each project has its own LLM, system prompt, tools, and configuration. Each model has a configurable context window with automatic chat memory management, and older messages are summarized rather than dropped.

Main features

  1. RAG uploads documents and queries them with LLM-powered retrieval.

  2. Block builds processing logic visually using a Blockly-based IDE, with no LLM required.

  3. Every system prompt change is automatically versioned.

  4. Every mutation across the platform (create, update, delete) is automatically logged, including who did what, when, and which resource was affected.

  5. Local and remote image generators are loaded dynamically.

  6. A built-in zero-shot text classifier supports multiple models.

Getting started

  1. Python 3.11 or higher is required.

  2. A Helm chart is provided.

From the README · bash
pip install restai-core
restai init      # Create database + admin user
restai migrate   # Run migrations
restai serve     # → http://localhost:9000/admin (admin / admin)
From the README · bash
restai serve -e .env -p 8080 -w 4
From the README · bash
docker run -p 9000:9000 apocas/restai:latest
# → http://localhost:9000/admin (admin / admin)
From the README · bash
make install    # Install deps, initialize DB, build frontend
make dev        # Development server with hot reload (port 9000)
make start      # Production server (4 workers, port 9000)

Who it is for

Inference provides direct LLM chat and completion. Direct Access uses LLMs, image generators, and audio transcription via OpenAI-compatible API endpoints, with no project required. A built-in evaluation system measures and tracks AI project quality over time.

Each team has its own users, admins, projects, and LLM/embedding access controls. Each team can customize the platform appearance for its members. Users can belong to multiple teams, each with optional custom branding.

The chat widget connects to a RESTai project, streams responses in real-time, and maintains conversation context. Local user accounts can be secured with TOTP-based two-factor authentication, compatible with Google Authenticator, Authy, and other authenticator apps. The Input Guard checks user questions before inference.

The Output Guard checks LLM responses after inference. Guards are regular RESTai projects. Rate limiting returns HTTP 429 when the limit is exceeded.

The maximum number of requests per minute is configured in the project edit page. RESTai automatically detects NVIDIA GPUs at startup. The listed databases are SQLite (default), PostgreSQL, and MySQL.

The project is licensed under the Apache License, Version 2.0. RESTai's Knowledge Base Sync keeps a project's RAG knowledge base up to date by syncing from external sources on a schedule, with configurable intervals from 15 minutes to 24 hours.

Sources

This page is written from the official sources below. Open them to check the original text.

Verified ·

This page is written by AI from the official sources above and checked automatically. This site has not tested it. Found a mistake? Tell us.