Deploy and Host OpenMAIC — Multi-Agent AI Classroom

Deploy OpenMAIC on Railway with one click. OpenMAIC (from Tsinghua's MAIC team) is an open-source multi-agent learning platform: describe a topic or upload your materials, and an AI teacher together with a classroom of agent students builds a full lesson — slides, quizzes, interactive HTML simulations, and project-based learning — in minutes.

This template deploys a fork of OpenMAIC tuned for Railway, with a Dockerfile fixed for Railway's builder (upstream cache-mount directives removed) so the image builds cleanly from source.

About Hosting

Hosting OpenMAIC on Railway gives you a private, always-on AI classroom with zero infrastructure work. Railway builds the Docker image from your connected GitHub repository, provides HTTPS automatically, and keeps a public URL live. Self-hosting means your lesson content, uploaded materials, and API keys stay under your control instead of a third-party SaaS. Because the service builds from a fork you own, you can add features or pull upstream changes and every push redeploys automatically.

Why Deploy

One-click lesson generation from a topic or your own documents
Multi-agent classroom: AI teachers and peer agents lecture, discuss, and answer in real time
Rich scenes: slides, quizzes, interactive simulations, project-based learning
Whiteboard and text-to-speech: agents draw diagrams and explain out loud
Export lessons as editable .pptx or interactive .html
Works with 20+ LLM providers — bring your own API key (OpenAI, Anthropic, Google, DeepSeek, Qwen, GLM, OpenRouter, and more)

Common Use Cases

Personal learning coach for any topic, on your own infrastructure
Study-group server: generate lessons once, share the URL with classmates
Corporate training: turn internal documents into interactive courseware
Tutoring prep: convert curricula into quizzes and simulations in minutes
Safe environment for kids — self-hosted means you control model choice and access

Dependencies for Deployment

Deployment Dependencies

At least one LLM provider API key (e.g. OPENAI_API_KEY or ANTHROPIC_API_KEY) — all other providers are optional
Optional: DEFAULT_MODEL to pin which model the classroom uses
Optional extras, each independently enabled by its own key: TTS (ElevenLabs, Azure, Qwen), image generation, video generation, web search (Tavily, Exa, Brave), and PDF parsing (MinerU)
No database required — runs standalone out of the box