A Production-Ready Model Context Protocol (MCP) Server for AI-Assisted Software Development

Empower AI assistants like ChatGPT and Claude with intelligent code understanding, multi-project workspaces, semantic search, Git integration, architecture analysis, and safe code editing.

---

✨ Overview

DevForge-MCP is a production-ready Model Context Protocol (MCP) server that enables AI coding assistants to intelligently understand, analyze, review, edit, and navigate software projects.

Unlike a traditional file server, DevForge-MCP builds an intelligent representation of your codebase, detects project architecture, manages multiple workspaces, integrates with Git, performs semantic code search, and exposes 44 developer tools through the latest Streamable HTTP MCP transport.

It is designed to work with modern AI assistants that support MCP, providing them with rich contextual understanding of your projects.

---

✨ Key Features

🧠 AI Code Intelligence

• Architecture detection
• Semantic code search
• Symbol lookup
• Dependency graph generation
• Context-aware navigation
• Duplicate code detection

📁 Workspace Management

• Multi-project support
• Project switching
• Automatic indexing
• Persistent project memory
• Incremental cache rebuilding

📸 Dashboard

✏ Safe Code Editing

• Patch files
• Replace methods
• Insert methods/classes
• Rename symbols
• Delete methods
• Multi-file editing

🌿 Git Integration

• Git status
• Commit changes
• Branch management
• Checkout branches
• Git history
• Diff inspection

🔍 Code Review

• Code smell detection
• Architecture review
• Refactoring suggestions
• Build error explanation
• Project review

⚡ Modern MCP

• Streamable HTTP
• JSON-RPC 2.0
• Stateless transport
• OpenAI Secure Tunnel compatible
• Latest MCP SDK

---

🏗 Architecture

                        AI Assistant
                     (ChatGPT / Claude)
                               │
                               ▼
                 Model Context Protocol (MCP)
                               │
                     Streamable HTTP Transport
                               │
                               ▼
                       DevForge-MCP Server
                               │
      ┌───────────────┬───────────────┬───────────────┐
      │               │               │
      ▼               ▼               ▼
 Code Intelligence    Git Engine    Workspace Manager
      │               │               │
      └───────────────┼───────────────┘
                      ▼
               Local Project Files

---

🚀 Core Capabilities

• Intelligent codebase indexing
• Multi-project workspaces
• Architecture detection
• Semantic code search
• Safe editing operations
• Git automation
• Build automation
• Code review
• Security inspection
• Project memory
• Dashboard
• Status API

---

📦 Technology Stack

• Node.js
• Express.js
• JavaScript
• Model Context Protocol SDK (v1.29+)
• Streamable HTTP
• Docker
• Railway
• Render

---

📂 Project Structure

DevForge-MCP
│
├── lib/
│   ├── architecture.js
│   ├── buildTools.js
│   ├── cache.js
│   ├── codeIntel.js
│   ├── editing.js
│   ├── fsTools.js
│   ├── gitTools.js
│   ├── indexer.js
│   ├── logger.js
│   ├── reviewTools.js
│   ├── security.js
│   └── state.js
│
├── mcp/
│   ├── handlers.js
│   ├── server.js
│   ├── tools.js
│   └── transport.js
│
├── public/
│
├── tests/
│
├── server.js
├── package.json
├── Dockerfile
├── README.md

---

🚀 Getting Started

Clone Repository

git clone https://github.com/Sanket2329/DevForge-MCP.git

cd DevForge-MCP

---

Install Dependencies

npm install

---

Start Server

npm start

---

Server Endpoints

Endpoint	Description
/mcp	MCP Endpoint
/dashboard	Web Dashboard
/status	Server Status

---

⚙ Environment Variables

Variable	Default	Purpose
PROJECT_PATH	Sample project	Initial project to index
WORKSPACE_ROOT	Parent directory	Workspace containing projects
PORT	3000	HTTP server port
MCP_API_KEY	unset	Protects REST API endpoints
LOG_LEVEL	info	Logging level

---

🧰 Developer Tool Inventory

Workspace

• list_projects
• select_project
• get_current_project
• refresh_index

Code Navigation

• get_project_files
• get_file_content
• get_full_codebase
• search_codebase
• get_architecture
• get_dependency_graph
• get_context

Safe Editing

• write_file
• write_multiple_files
• patch_file
• replace_method
• insert_method
• insert_class
• rename_symbol
• delete_method

Code Intelligence

• find_symbol
• find_class
• find_method
• find_interface
• find_references
• find_unused_files
• find_duplicate_code

Git

• git_status
• git_diff
• git_log
• git_checkout
• create_branch
• commit_changes
• show_untracked_files

Review

• review_file
• review_project
• find_code_smells
• review_architecture
• suggest_refactoring
• explain_build_error

Memory

• add_conversation_turn
• get_conversation_history
• add_project_note
• get_project_memory

---

🧪 Testing

The project has been validated with:

✅ SDK Import Tests

✅ Streamable HTTP Transport

✅ JSON-RPC Request Handling

✅ MCP Tool Registration

✅ Tool Discovery (44 Tools)

✅ Integration Tests

✅ OpenAI Secure Tunnel

✅ Workspace Switching

✅ Codebase Indexing

---

🔄 Streamable HTTP Migration

DevForge-MCP has been migrated from the legacy Server-Sent Events (SSE) transport to the latest Streamable HTTP transport.

Migration Highlights

• Removed legacy SSE transport
• Stateless request handling
• JSON-RPC 2.0 compliance
• Modern MCP SDK compatibility
• OpenAI Secure Tunnel support
• Improved scalability

---

🌐 Deployment

DevForge-MCP supports deployment using:

• Docker
• Railway
• Render

Deployment configuration files included:

• Dockerfile
• railway.toml
• render.yaml

For deployment instructions, see:

DEPLOYMENT.md

---

🔌 Connecting AI Assistants

MCP Endpoint

https://your-domain.com/mcp

Compatible with:

• ChatGPT (Developer Mode)
• Claude Desktop
• MCP-compatible clients

Authentication:

No Authentication

unless configured otherwise.

---

⚠ Known Limitations

• Architecture detection is heuristic-based.
• Code review is pattern-based rather than compiler-backed.
• Duplicate code detection uses hash comparison instead of semantic analysis.
• Editing operations rely on syntax heuristics for different languages.

---

📈 Roadmap

• Vector Search
• Embedding Support
• RAG Integration
• Plugin System
• Authentication
• GitHub Actions CI/CD
• Performance Metrics
• Language Server Protocol Integration

---

🤝 Contributing

Contributions are welcome.

1. Fork the repository.
2. Create a feature branch.
3. Commit your changes.
4. Submit a Pull Request.

---

📄 License

This project is licensed under the MIT License.

---

👨‍💻 Author

Sanket Shakya

• GitHub: https://github.com/Sanket2329

---

⭐ Support

If you find DevForge-MCP useful, consider giving the repository a ⭐ Star on GitHub.

It helps others discover the project and supports future development.