Codebase Insights MCP Server

An MCP (Model Context Protocol) server that provides comprehensive codebase analysis, API documentation, Postman collection generation, and business insights. Analyzes API codebases to generate Postman collections, Product Owner reports, and detailed code insights. Supports multiple frameworks including FastAPI, Spring Boot, Flask, Express, and any codebase with OpenAPI/Swagger specifications.

Features

• 🚀 Multi-Framework Support: FastAPI, Spring Boot, Flask, Express, NestJS

• 📋 OpenAPI/Swagger First: Automatically detects and uses OpenAPI specs when available

• 🔍 Smart Code Analysis: Falls back to AST/pattern matching when no spec is found

• 🔐 Authentication Support: Handles Bearer, Basic, and API Key authentication

• 📁 Organized Output: Groups endpoints by tags or path segments

• 🎯 Accurate Detection: Extracts request bodies, parameters, and response examples

• 🤖 OpenAI ChatGPT Integration: SSE transport support for ChatGPT Connectors and Deep Research

• 🔌 Multiple Transport Modes: stdio (local), HTTP (testing), SSE (OpenAI integration)

Installation

For End Users (via uvx)

The easiest way to use this MCP server is with uvx:

For Development

Configuration

Claude Desktop Configuration

Add this to your Claude Desktop configuration file:

For uvx installation (recommended):

For local development:

Environment Variables

• BITBUCKET_EMAIL: Your Bitbucket username (not email)

• BITBUCKET_API_TOKEN: Bitbucket app password with repository read access

• GITHUB_TOKEN: GitHub personal access token (for private repos)

• output_directory: Where to save generated Postman collections (default: current directory)

Usage

Once configured, you can use the following commands in Claude:

The server will:

1. Clone the repository

2. Detect the framework and analyze the codebase

3. Extract all API endpoints with their parameters and examples

4. Generate a Postman v2.1 collection file

5. Save it to your output directory

Supported Frameworks

With Full Support

• FastAPI (Python) - Full OpenAPI integration

• Spring Boot (Java) - Annotation-based detection

• Express (Node.js) - Route pattern matching

• Flask (Python) - Decorator-based detection

• Django REST (Python) - ViewSet and path detection

Coming Soon

• NestJS (TypeScript)

• Ruby on Rails

• ASP.NET Core

Publishing Updates to PyPI

One-Time Setup

Publishing Updates Workflow

1. Update Version

   # Update version in pyproject.toml poetry version patch # for bug fixes (0.1.0 -> 0.1.1) poetry version minor # for new features (0.1.0 -> 0.2.0) poetry version major # for breaking changes (0.1.0 -> 1.0.0)

2. Build Package

   poetry build

3. Publish to PyPI

   poetry publish

4. Test Installation

   # Test the published package uvx --reinstall codebase-insights-mcp

Automated Version Workflow

Users Update with uvx

After publishing, users can update with:

Development

Running Tests

Code Quality

Local Testing with MCP Inspector

Option 1: Test Published Version (HTTP)

Option 2: Test Development Version (HTTP)

Testing SSE Transport (OpenAI ChatGPT Integration)

The SSE transport is required for OpenAI ChatGPT Connectors and Deep Research integration.

Local SSE Testing

Option 1: Test Published Version with SSE

Option 2: Test Development Version with SSE

Testing with OpenAI API

Once your SSE server is running and publicly accessible, test it with OpenAI:

Available Tools for OpenAI Integration

The following tools are available when using SSE transport with OpenAI:

1. search - Search for repositories and code (required by OpenAI ChatGPT Connectors)

   { "query": "FastAPI authentication" }

2. fetch - Fetch full document content by ID (required by OpenAI ChatGPT Connectors)

   { "id": "doc-123" }

3. generate_collection - Generate Postman collections from repositories

   { "repo_url": "https://github.com/user/repo.git" }

4. generate_product_owner_overview - Generate business analysis reports

   { "repo_url": "https://github.com/user/repo.git" }

5. analyze_repository_for_llm - Clone and return code for LLM analysis

   { "repo_url": "https://github.com/user/repo.git", "max_files": 50 }

Transport Modes Comparison

When to use each transport:

• stdio: Default for local MCP clients like Claude Desktop

• HTTP: Best for testing with MCP Inspector during development

• SSE: Required for OpenAI ChatGPT Connectors and Deep Research features

Using the Test Scripts

Interactive MCP Inspector Testing:

Automated HTTP API Testing:

Testing with MCP Inspector

1. Open MCP Inspector or run locally

2. Enter URL: http://localhost:8000/mcp

3. Click Connect

4. Test tools:

   • generate_collection with: {"repo_url": "https://bitbucket.org/tymerepos/tb-payshap-svc.git"}

   • get_server_info with: {}

STDIO

• Command /bin/bash

• Arguments -c "cd '/Users/lukelanterme/Documents/Code/Personal/AI/Projects/codebase-insights-mcp' && poetry run codebase-insights-mcp"

export BITBUCKET_EMAIL="your-username" export BITBUCKET_API_TOKEN="your-app-password" export output_directory="/tmp/"

Architecture

The server follows clean architecture principles:

• Models: Pydantic models for API endpoints and Postman collections

• Analyzers: Framework-specific endpoint extraction logic

• Generators: Postman collection generation from API models

• Clients: Git repository access with authentication support

Contributing

1. Fork the repository

2. Create a feature branch (git checkout -b feature/amazing-feature)

3. Commit your changes (git commit -m 'Add amazing feature')

4. Push to the branch (git push origin feature/amazing-feature)

5. Open a Pull Request

License

MIT License - see LICENSE file for details

Acknowledgments

• Built with FastMCP framework

• Inspired by API development workflows and the need for automation