About this project

OpenMAIC (Open Multi-Agent Interactive Classroom) is an open-source AI platform for generating interactive, multi-agent classroom experiences from a topic or uploaded materials. It is built with Next.js, React, TypeScript, LangGraph and Tailwind CSS, and is released under the MIT license. What it does - One-click lesson generation: describe a topic or attach documents, audio or video, and the system builds a full lesson. - Multi-agent classroom: AI teachers and AI classmates lecture, discuss, draw on a whiteboard and speak via TTS. - Scene types include slides, quizzes, interactive HTML simulations, 3D visualization, games, mind maps, online programming and project-based learning (PBL) activities. - Export options include editable .pptx slides and interactive .html pages, plus MP4 video export and offline classroom ZIP export/import. Agent workbench and skills - A SKILL.md-format skill package lets agent workbenches such as OpenClaw, Codex, DeepSeek and WorkBuddy generate classrooms from messaging apps or an IDE. - A Pro workbench provides a chat-first workspace that plans, builds and revises courses, with durable server-backed sessions that can be cancelled, resumed and steered. - Session materials can be uploaded or pulled from web search. Providers and local AI - Supports many LLM providers including OpenAI, Azure OpenAI, Anthropic, Amazon Bedrock, Google Gemini, DeepSeek, Qwen, Kimi, MiniMax, Grok, OpenRouter, Doubao, Tencent Hunyuan, Xiaomi MiMo, GLM and any OpenAI-compatible API. - Local options include Ollama, Lemonade (LLM, image, TTS, ASR) and FunASR for local speech recognition. - Separate providers can be configured for TTS, ASR, image and video generation. Deployment and persistence - Quick start requires Node.js >= 22.19 and pnpm >= 10; configuration is via .env.local or server-providers.yml. - Deployment options include local dev, production build, Vercel and Docker Compose, with optional build arguments for mirror acceleration. - An optional server-persistence profile runs the app with PostgreSQL for server-backed course documents and runtime sessions; browser-scoped data remains in the browser. - An optional ACCESS_CODE provides a site-level password for shared deployments. - Asset storage includes a registry, quota enforcement and an offline collector for unreferenced bytes. Documentation and community - English and Simplified Chinese user guides, a live demo, a JCST paper reference, Discord and Feishu communities, and a changelog covering releases from v0.1.0 through v1.0.0.