Graphify-Labs/graphify
Graphify transforms codebases, documentation, and other files into queryable knowledge graphs, enhancing developer productivity with local AST parsing and detailed edge explanations.
Awesome MCP › Software Development
haasonsaas/deep-code-reasoning-mcp
Deep Code Reasoning MCP is a TypeScript Model Context Protocol server that lets an MCP client such as Claude Code hand analysis tasks to Google Gemini 2.5 Pro Preview. The design treats the two models as heterogeneous services: the client keeps local context, incremental patches and CLI-native refactoring, while the server takes work that benefits from Gemini's one-million-token context window and code execution, such as sweeping large log and trace dumps, correlating failures across microservices, and running iterative hypothesis tests. The tools cover execution flow tracing, which follows data flow and state transformations rather than call graphs alone; cross-system impact analysis, which models how a change propagates across service boundaries; performance modelling for N+1 query patterns, memory leaks and algorithmic bottlenecks; and hypothesis testing that checks theories about code behaviour against evidence. A conversational set of tools named start_conversation, continue_conversation and finalize_conversation keeps a multi-turn session between the two models, carrying the caller's attempted approaches, partial findings, a description of where it got stuck, and a code scope of files, entry points and service names; the session ends with a structured analysis result. Tool parameters use snake_case names and are validated with Zod schemas, with the TypeScript definitions kept in the repository. The server is built on the official MCP SDK together with Google's generative AI SDK, requires Node.js 18 or later, and reads a Gemini API key from the GEMINI_API_KEY environment variable. Installation is a git clone, an npm install and a build, after which the server is registered in a Claude Desktop or Cursor MCP configuration as a node process. It targets developers who already work through an MCP-capable coding assistant and run into context or reasoning limits on large distributed codebases. The project is MIT licensed.
https://github.com/haasonsaas/deep-code-reasoning-mcp
Also in Software Development
upstash/context7
Context7 MCP Server is an MCP server that delivers up-to-date, version-specific code documentation and examples to enhance LLMs and AI code editors, ensuring accurate and current code generation.
DeusData/codebase-memory-mcp
High-performance Model Context Protocol (MCP) server that indexes codebases into a persistent knowledge graph for AI coding agents, supporting 158 languages, lightning-fast queries, and various dev...
github/github-mcp-server
GitHub MCP Server is an official Model Context Protocol server that integrates with GitHub APIs to enable advanced automation and AI-powered interactions with GitHub repositories and workflows.
tirth8205/code-review-graph
Code-review-graph is a local-first code intelligence solution that builds a persistent structural map of your codebase to provide precise context to AI coding tools, significantly reducing token us...
oraios/serena
Serena is a free and open-source coding agent toolkit that uses the Model Context Protocol to enable LLMs to autonomously perform semantic code retrieval and editing with IDE-like precision on complex codebases.
eyaltoledano/claude-task-master
Claude Task Master is an AI-powered task management system designed for AI-driven development that integrates seamlessly with various editors via the Model Control Protocol (MCP).
mksglu/context-mode
Context Mode is an MCP server designed to optimize context window usage for AI coding agents by sandboxing tool output, tracking session continuity with SQLite, and promoting LLMs as code generator...