Looma is a web application that lets you generate stunning animations just by describing them in natural language. Powered by Manim under the hood, Looma translates your prompt into Python animation code, renders it using a backend worker, and delivers a playable video — all in one seamless flow.
Built with Next.js, FastAPI, and OpenAI, Looma is your creative canvas for math, science, and educational visualizations — no coding required.
- 🧠 Prompt-to-Animation: Describe your animation and get a rendered video
- 🎬 Manim Integration: Uses Manim to create mathematically accurate visuals
- ⚡ FastAPI Worker: A Python-based worker renders Manim code securely
- 🔐 Google OAuth: Authenticate via your Google account
- 🌐 Next.js Frontend: Smooth UI built with React and deployed via Next.js
Looma/
├── src/ # Next.js application (frontend)
├── worker/ # Manim worker service (FastAPI)
├── .env # Environment configuration
├── pnpm-workspace.yaml # Monorepo workspace config
└── README.md # You're reading it :)
Create a .env file in the root with the following:
# Database
DATABASE_URL=
DIRECT_URL=
# Google OAuth
GOOGLE_CLIENT_ID=
GOOGLE_CLIENT_SECRET=
# Gemini API
GOOGLE_GENERATIVE_AI_API_KEY=
# OpenAI API
OPENAI_API_KEY=
# Auth
NEXTAUTH_SECRET=secret
# Manim Worker URL
WORKER_URL="http://0.0.0.0:8000"- Frontend: Next.js 14, React, Tailwind CSS, NextAuth.js, Prisma
- Backend Worker: FastAPI, Manim (Python)
- Authentication: Google OAuth
- AI Integration: OpenAI (GPT for prompt → code generation)
- Package Manager:
pnpm(monorepo with workspace support)
git clone https://github.com/yourusername/looma.git
cd looma
pnpm installCreate a .env file at the root using the example above.
Follow the instruction provided in the folder
pnpm dev"Make a sqaure"
- Add a code editor
- User gallery and saved animations
- More animation styles and themes
- Real-time preview and scrubber
Pull requests are welcome! Please open an issue first to discuss what you’d like to change.