Atlas MCP Server

The Agentic AI Pipeline for your IDE

Advanced AI/ML-powered development tools with RAG search, ML bug prediction, intelligent code migration, and comprehensive test generation.

Installation • Features • Tools • Usage • Documentation

Overview

Atlas MCP Server gives your IDE "Agentic Superpowers" through a full analysis pipeline:

Intent → Context → Decomposition → Variants → Critique → Optimization

Works natively with Cursor, Windsurf, Claude Desktop, GitHub Copilot, and VS Code.

✨ What's New in v1.0.26

🎨 4 Powerful Frontend Developer Tools (NEW!)

These tools solve real-time frontend development problems with production-ready solutions:

Previous Releases

v1.0.25 - 5 Senior Developer Tools

Previous Releases (v1.0.23)

Advanced AI/ML Tools (4)

Core Professional Tools (13)

✨ What's New in v1.0.23

🎯 Key Features

🤖 AI/ML Capabilities

• RAG Search - Semantic code understanding with knowledge graphs
• ML Prediction - Bug & performance forecasting (70-85% accuracy)
• Code Migration - 12 intelligent migration types
• Test Generation - Advanced test suites with edge cases

⚡ Professional Tools

• Full Pipeline - Intent → Context → Variants → Optimize
• Quality Analysis - Code review, security, complexity
• Performance - Profiling, bottleneck detection
• Automation - Tests, docs, refactoring

🌟 Why Atlas?

• Zero Config: No API keys required - uses your IDE's built-in AI (Copilot/Cursor)
• 23 Professional Tools: 13 core + 4 advanced AI/ML + 5 senior developer + 1 UI/UX designer
• Agentic Workflow: DAG-based task decomposition
• Context Aware: Project structure, dependencies, git history analysis
• High Performance: LRU caching, request deduplication, parallel execution

Installation (1-Click Setup)

1. Install Globally

Open your terminal and run the following command to install the package globally via NPM:

npm install -g atlas-pipeline-mcp

2. Run Auto-Setup

Run the setup command to automatically configure your IDE (works for Cursor and VS Code):

atlas-mcp-setup

3. Restart IDE

Restart your editor. You should see the Atlas server connected in your MCP settings.

🛠️ All Available Tools (27 Total)

🔥 Advanced AI/ML Tools

⚡ Core Pipeline & Analysis Tools

🎯 Quality & Optimization Tools

🧠 Advanced Senior Developer Tools (NEW!)

🎨 Design & UI/UX Tools

🚀 Frontend Developer Tools (NEW in v1.0.26!)

🚀 Development Productivity Tools

🔧 Utility Tools

How to Use (Cheat Sheet)

Once installed, simply chat with your AI Assistant (Copilot Chat or Cursor Chat). The server automatically activates based on your intent.

Advanced AI/ML Tools

| Goal | What to Ask | Tool Used | | :--- | :--- | :--- | | Semantic Code Search | "Search for authentication code using RAG" | atlas_rag_search | | Predict Bugs | "Predict bug probability in payment-processor.ts" | atlas_ml_predict | | Migrate Codebase | "Migrate this file from JavaScript to TypeScript" | atlas_migrate | | Advanced Testing | "Generate comprehensive tests with edge cases" | atlas_testgen_advanced |

Professional Development Tools

| Goal | What to Ask | Tool Used | | :--- | :--- | :--- | | Fix a complex file | "Run the pipeline on utils.ts to refactor it." | atlas_pipeline | | Refactor Code | "Refactor this code and reduce complexity" | atlas_refactor | | Performance Analysis | "Profile this code for performance bottlenecks" | atlas_profiler | | Code Review | "Review this PR for quality and security" | atlas_review | | Check Dependencies | "Analyze dependencies and find unused packages" | atlas_dependencies | | Plan a feature | "Decompose the task of adding JWT auth." | atlas_decompose | | Explore ideas | "Generate 3 variants for this button component." | atlas_variants | | Review Code | "Critique this code for security issues." | atlas_critique | | Polish Code | "Optimize this function based on best practices." | atlas_optimize | | Security Scan | "Scan this code for security vulnerabilities." | atlas_security | | Generate Tests | "Generate tests for this authentication module." | atlas_test | | Add Documentation | "Document this API with examples." | atlas_docs | | Understand Code | "Explain how this algorithm works." | atlas_explain | | Debug Error | "Debug this TypeError, here's the stack trace..." | atlas_debug | | Complex Reasoning | "Think through how to design this system." | atlas_think | | Project Context | "Analyze the project context and dependencies." | atlas_context |

Advanced Integration (Optional)

Using Local LLMs

If you prefer running models locally (e.g. Ollama) or want to use your own API keys instead of your IDE's subscription, you can manually configure the server in your settings:

"atlas": {
  "command": "npx",
  "args": ["-y", "atlas-pipeline-mcp"],
  "env": {
    "OLLAMA_BASE_URL": "http://localhost:11434",
    "OPENAI_API_KEY": "sk-...",
    "ANTHROPIC_API_KEY": "sk-..."
  }
}

Note: If no keys are provided, Atlas defaults to Client Sampling mode, delegating generation to your IDE.

Contributing

We welcome contributions to improve the Atlas pipeline.

• Repository: github.com/IamNishant51/atlas-mcp-server
• Issues: Report bugs on GitHub.

Built by Nishant Unavane