Overview
OpenMAIC (Open Multi-Agent Interactive Classroom) is an open-source app from THU-MAIC, a research team at Tsinghua University, that turns a topic or an uploaded document into a lesson you can sit through. A group of agents plans the curriculum, writes the material and then delivers it: slides, quizzes, interactive HTML simulations and project-based activities, with text-to-speech narration and a whiteboard you can draw on while the lesson runs.
It is a multi-agent system rather than a single prompt-to-slides generator. Separate agents take the teacher and peer roles inside the classroom, and a Pro-mode agent workbench lets you chat with an agent that plans the curriculum and then builds and revises every page. Lessons export to PowerPoint, interactive HTML or MP4.
OpenMAIC is provider-neutral. It runs against OpenAI, Anthropic, Google and other hosted APIs through your own key, and can also point at a local model server such as Ollama or Lemonade. Persistence is pluggable, with an optional PostgreSQL backend for server-side storage, and the interface ships in twelve locales.
What it does
- One-click lesson generation from a topic or an uploaded document
- Multi-agent classroom with agent teachers and peer roles rather than a single generator
- Mixed scene types — slides, quizzes, interactive HTML simulations and project-based activities
- Text-to-speech narration plus a shared whiteboard for drawing during a lesson
- Agent workbench (Pro mode) that plans a curriculum and revises pages through chat
- Export to .pptx, interactive HTML or MP4
- Bring your own key across OpenAI, Anthropic, Google and other providers, or point at a local model server
- Twelve-locale interface and optional PostgreSQL-backed persistence
Getting started
OpenMAIC is a TypeScript app you run locally with pnpm. You supply your own model API key.
Clone and install
OpenMAIC 1.0.1 requires Node 22.19 or newer.
git clone https://github.com/THU-MAIC/OpenMAIC.git
cd OpenMAIC
pnpm installAdd a model API key
Copy the example environment file and set the key for whichever provider you want the agents to call.
cp .env.example .env.local
# then edit .env.local and add your OpenAI, Anthropic or Google keyRun the app
Start the dev server and open the classroom in a browser.
pnpm dev
# http://localhost:3000Pointing at a local model server
Since 1.0.1, a client-supplied base URL on a loopback or private address is rejected in development builds too. Opt back in if you are running a local Ollama or Lemonade endpoint.
ALLOW_LOCAL_NETWORKS=true pnpm devCommands and code are distilled from the project's own documentation — always check the official repo for the latest.
When to use it
- Turn a paper, a spec or a set of notes into a lesson you can actually sit through instead of skim
- Generate slide decks and quizzes for a course and export them to PowerPoint
- Give a self-taught learner an agent teacher plus peer agents to ask questions alongside
- Build interactive simulations and programming exercises for a topic without hand-authoring the HTML
How OpenMAIC compares
OpenMAIC alongside other open-source multi-agent systems tools AI/TLDR tracks, ranked by GitHub stars.
| Tool | Stars | What it does |
|---|---|---|
| TradingAgents | ★ 103k | LangGraph framework that mirrors a trading firm: LLM analyst, bull/bear researcher, trader and risk-management agents debate before a decision. For research, not advice. |
| Ruflo | ★ 71k | Agent meta-harness that wraps Claude Code and Codex with 100+ specialized agents, swarm coordination, vector memory, background workers and cross-machine agent federation. |
| MetaGPT | ★ 70.2k | A multi-agent framework that models a software company, assigning roles like product manager, architect, and engineer to generate code from a single prompt. |
| AutoGen | ★ 60.8k | Microsoft Research's framework for building applications where multiple agents converse with each other and with tools to solve tasks. |
| CrewAI | ★ 58.2k | A framework for assembling teams ('crews') of role-playing agents that divide tasks and collaborate to complete a goal. |
| AgentScope | ★ 30.9k | A framework for building multi-agent applications with message passing, visual debugging tools, and distributed execution. |
| OpenAI Agents SDK | ★ 29.2k | OpenAI's lightweight Python SDK for building multi-agent workflows using explicit handoffs, tools, and guardrails. |
| OpenMAIC | — | Turn a topic or a document into an agent-taught interactive lesson |