← Discover MCPs and Agents
s
MCPAI & MLGitHub

spark-stack

Spark Stack is an tool for building web applications through an AI-powered chat interface. Create quick MVPs and prototypes using natural language prompts.

Links

README

From the repo.

Spark Stack

Spark Stack is an tool for building web applications through an AI-powered chat interface. Create quick MVPs and prototypes using natural language prompts. [Blog Post]

Update: Jun 1, 2025: Took down the demo instance and have effectively archived the project. Afaik it still works but have moved on to other projects.

chrome_vMZlrhHm0u

Deploy on Railway

Features

  • 🤖 AI-powered code generation
  • ⚡️ Real-time development environment
  • 🎨 Multiple arbitrary starter templates (see /images)
  • 👥 Team collaboration and management
  • 📝 Git version control
  • 🔄 Live preview
  • 🧠 Chain-of-Thought reasoning for complex asks
  • 🔌 Support for OpenAI and Anthropic models
  • 📱 Multi-page app generation
  • 📸 Sketch and screenshot uploads
  • 🚀 Deployment to GitHub (+ Netlify, Vercel, etc)
  • 🌙 Dark mode support
  • 🔗 Share chats and projects publicly
  • 💾 Prompt caching for cheaper responses
  • 🖥️ Virtual browser logs and screenshots for debugging

Setup

Environment Configuration

See backend/config.py for the environment variables that are used to configure the app.

  • Requires modal account to be created and configured.
  • Requires AWS account and s3 bucket to be configured.

Development

  • cd frontend && npm install && npm run dev
  • cd backend && pip install -r requirements.txt && python main.py

Deployment

Railway (docker + postgres).

chrome_E9GXwtsE87

LoC History

This project was a pressure test for writing code quickly with Cursor so I thought it was interesting to graph how it was built.

screenshot

Red is my initial 2-day sprint to get an MVP (at this point it worked fully e2e but was a bit brittle). Dots are commits that I arbitrarily checkpointed as I was working on the project.

Collected info

  • 278 stars
  • 37 forks
  • Language: JavaScript
  • Source updated: 9/18/2026

Config for your environment

Replace {MCP_ENDPOINT_URL} with this MCP’s endpoint URL (from its repo or docs above). No API key — you connect directly.

Tool

OS

Config file: ~/.cursor/mcp.json

{
  "mcpServers": {
    "mcp-server": {
      "url": "{MCP_ENDPOINT_URL}"
    }
  }
}

Paste into mcpServers in the config file. Restart Cursor after saving.

If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.