Run multiple AI agents in parallel to prototype websites and apps
Participate at 1000prototypes.space
Instead of waiting for one AI assistant to build your idea, this tool lets you deploy multiple specialized agents that work on the same request simultaneously. You get different approaches and can pick the best parts from each.
I got tired of the back-and-forth with AI tools, I mostly want to shoot out multiple trys when building webinterface prototypes to choose a cool one. You describe something, get one interpretation, then spend forever trying to guide it toward what you actually wanted.
This approach is different: set up multiple agents with different specialties, give them all the same brief, and see what each one comes up with. One might nail the layout, another might have better functionality, and a third might surprise you with a creative approach you hadn't considered.
- Deploy multiple AI agents with custom prompt sequences
- Compare results side by side and see different interpretations
- Share agents you've created with the community
- Try different AI models to see which works best for different tasks
- Save prototypes you like to a gallery
The community has built agents for e-commerce sites, portfolios, SaaS landing pages, blogs, interactive experiences, and mobile apps. Check out 1000prototypes.space to see what's possible.
# Clone the repo
git clone https://github.com/DomEscobar/1000prototypes.git
cd creative-ai-architects
# Install dependencies
npm install
cd server && bun install
# Start everything
npm run dev:fullcd docker
docker-compose up -dWorks on Railway, AWS, Google Cloud, or wherever you can run Docker.
Each agent has a sequence of prompts that build on each other. The first prompt might analyze your request and plan the approach, the second might generate the HTML structure, and the third might add styling and interactions.
This is an open source project and contributions are welcome. Whether you want to:
- Create new agent templates for the community
- Improve the UI/UX of the platform
- Add new features to the backend
- Fix bugs or improve performance
- Write documentation or tutorials
Just open an issue or submit a PR. The codebase is fairly straightforward - React frontend with a Bun/Hono backend.
Frontend: React + TypeScript, Vite, Tailwind CSS + shadcn/ui
Backend: Bun + Hono with streaming support
Storage: JSON files (simple and effective for this use case)
Deployment: Docker + Nginx
- Issues/Features: Use GitHub issues for bugs or feature requests
- Discussions: GitHub discussions for questions and ideas
The goal is to build a useful tool for developers who want to prototype faster and explore different AI approaches to the same problem.
MIT - build whatever you want with it.