Atlas MCP Server

A powerful Model Context Protocol (MCP) server featuring a multi-stage AI pipeline for intelligent code generation. Works with any LLM provider (Ollama, OpenAI, Anthropic) and any MCP-compatible IDE (Cursor, GitHub Copilot, Claude Desktop, Windsurf, and more).

🚀 Features

7-Stage AI Pipeline: Intent → Context → Git → Decompose → Variants → Critique → Optimize
Multi-Provider Support: Automatically detects and uses Ollama, OpenAI, or Anthropic
IDE Agnostic: Works with any MCP-compatible editor
Git Integration: Analyzes repository history for context-aware code generation
Quality Assurance: Built-in critique and optimization stages

📦 Installation

From GitHub (Recommended)

# Clone the repository
git clone https://github.com/IamNishant51/atlas-mcp-server.git
cd atlas-mcp-server

# Install dependencies
npm install

# Build
npm run build

From npm

npm install -g atlas-pipeline-mcp

⚙️ Configuration

Environment Variables

Create a .env file or set environment variables:

# For Ollama (local, free)
OLLAMA_BASE_URL=http://localhost:11434

# For OpenAI
OPENAI_API_KEY=your-openai-key

# For Anthropic
ANTHROPIC_API_KEY=your-anthropic-key

The server auto-detects available providers in this order: Anthropic → OpenAI → Ollama

IDE Configuration

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "atlas": {
      "command": "node",
      "args": ["/path/to/atlas-mcp-server/dist/mcp.js"]
    }
  }
}

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "atlas": {
      "command": "node",
      "args": ["/path/to/atlas-mcp-server/dist/mcp.js"]
    }
  }
}

VS Code with GitHub Copilot

Add to .vscode/mcp.json:

{
  "servers": {
    "atlas": {
      "command": "node",
      "args": ["/path/to/atlas-mcp-server/dist/mcp.js"]
    }
  }
}

Windsurf

Add to ~/.windsurf/mcp.json:

{
  "mcpServers": {
    "atlas": {
      "command": "node",
      "args": ["/path/to/atlas-mcp-server/dist/mcp.js"]
    }
  }
}

🛠️ Available Tools

Tool	Description
`atlas_intent`	Analyze user intent and extract requirements
`atlas_context`	Gather project context (files, dependencies, patterns)
`atlas_git`	Analyze git history for coding patterns
`atlas_decompose`	Break down tasks into subtasks
`atlas_variants`	Generate multiple code implementation variants
`atlas_critique`	Evaluate code quality and provide feedback
`atlas_optimize`	Optimize code based on critique feedback
`atlas_pipeline`	Run the complete 7-stage pipeline
`atlas_providers`	List available LLM providers

📊 Pipeline Stages

Intent - Understands what you want to build
Context - Gathers project structure, dependencies, and patterns
Git - Analyzes commit history for coding conventions
Decompose - Breaks complex tasks into manageable subtasks
Variants - Generates multiple implementation approaches
Critique - Evaluates each variant for quality and correctness
Optimize - Refines the best variant based on feedback

🏃 Running

MCP Mode (for IDEs)

npm run start:mcp
# or
node dist/mcp.js

HTTP Server Mode

npm start
# or
node dist/server.js

The HTTP server runs on port 3000 with these endpoints:

GET /health - Health check
GET /api/info - Server information
GET /api/models - Available LLM models
POST /api/pipeline - Run the full pipeline

🤝 Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

📄 License

MIT License - see LICENSE for details.

🔗 Links

GitHub: https://github.com/IamNishant51/atlas-mcp-server
MCP Protocol: https://modelcontextprotocol.io

atlas-pipeline

Quick Install