MCP Servers: Developer Tools

everything91.2k★

The Everything server is a test rig, not a tool. It exercises the MCP features a client has to handle: tools, prompts, resources, sampling, elicitation, logging, progress notifications and argument completions, so client builders can check their work against something known. One npx command runs it over stdio, SSE or streamable HTTP. There is no service behind it and nothing to pay for. The MCP steering group maintains it as a reference implementation. Building or debugging an MCP client? Install it. Everyone else can skip it, and nobody should run it where the environment holds secrets: its get-env tool returns every environment variable to the model.

memory91.2k★

Memory gives your AI assistant a notebook that survives between chats. It stores a small knowledge graph of entities, the relations between them and facts attached to each, then lets the model search and open those nodes later. Install it with npx or Docker. Everything lives in one local JSONL file, and MEMORY_FILE_PATH puts that file somewhere you control and back up. No accounts, no service costs. The MCP steering group maintains it as a basic reference implementation. It works well for personal context like preferences and project facts. Need semantic search, sync across machines or memory shared by a team? That is a different class of tool.

filesystem91.2k★

Filesystem lets an AI assistant read and change files, but only inside directories you allow. It covers reading text and media, writing, pattern-based edits with a dry-run preview, creating and listing directories, moving files, search and directory trees. Set the allowed paths as arguments, or let a client that supports MCP Roots update them at runtime. It starts with npx, or in Docker with mounts you can mark read-only. Local and free. The MCP steering group maintains it as a reference implementation, not a production product. For local work it is the obvious first install. Point it at a narrow folder, because write_file overwrites existing files.

sequentialthinking91.2k★

Sequential Thinking gives the model a structured scratchpad. Instead of answering in one shot, it records numbered thoughts, revises earlier ones, branches into alternatives and adjusts how many steps it expects to need. One npx or Docker entry sets it up, and DISABLE_THOUGHT_LOGGING silences the thought log. The server is local and free, but every extra thought is another tool call your model has to generate, so answers cost more tokens. The MCP steering group maintains it as a reference implementation. It earns its place on migrations, production debugging and architecture tradeoffs where course correction matters. Models with strong built-in reasoning get less out of it.

fetch91.2k★

Fetch hands your AI assistant a URL reader. The model asks for a page, and the server returns it converted to markdown, in chunks the model can page through when the content runs long. Setup is one uvx mcp-server-fetch entry in your config, or a Docker image. It obeys robots.txt for model-initiated requests unless you pass --ignore-robots-txt, and it takes a custom user agent and a proxy. No API keys, no service fees. The MCP steering group maintains it as a reference implementation, and its README carries a caution worth reading twice: it can reach local and internal IP addresses. Fine on a laptop. On any network with internal services, sandbox it first.

git91.2k★

The Git server lets an AI assistant work a local repository directly: status, staged and unstaged diffs, diffs against other branches, log with date filters, show, branch listing and creation, checkout, add, reset and commit. Run it with uvx mcp-server-git against any local repo. No GitHub account, no API keys, no costs. It has no push, pull or other remote operations. The MCP steering group maintains it, and its README says it is in early development with tools subject to change. Install it when your client has no Git tooling of its own. Coding agents that already run git through a shell get little from it.

time91.2k★

Time fixes a small but real gap: models do not know what time it is. It exposes two tools, one for the current time in any IANA timezone and one for converting a clock time between zones, and detects your system timezone automatically. Run it with uvx mcp-server-time or Docker. Override the detected zone with --local-timezone if you need to. Nothing to pay. The MCP steering group maintains it as a reference implementation. It does one job: conversion takes an HH:MM time, not a full date, and there is no calendar logic. Cheap to add when deadlines or cross-timezone coordination come up in your prompts.

github-mcp-server33.6k★

GitHub's official MCP server gives your AI assistant direct access to repos, issues, PRs, code search, and file operations. No scraping, no workarounds. You talk to your assistant, it talks to GitHub. Setup is one config block plus a personal access token. Maintained by GitHub themselves, so it stays current with API changes. Covers the operations you actually use daily: creating branches, reviewing diffs, searching code across repos, managing issues. The scope is broad enough that it replaces most of your tab-switching. Worth installing immediately if you use GitHub. This is the one MCP server that pays for itself on day one. The catch: it can only do what the GitHub API allows, so anything requiring the web UI (like editing Actions workflows visually) is still manual.

Figma-Context-MCP16.0k★

Connects your AI coding agent directly to Figma designs. Instead of screenshots or manual descriptions, the model reads your actual design files, including layout structure, styles, and component hierarchies. It bridges the gap between design and implementation. You get read access to Figma files, frames, and components with structured context the model can act on. Requires a Figma access token. The MCP server is free, Figma itself requires a paid plan for most team usage. Maintained by the community (GLips). This is the go-to Figma integration for AI coding workflows. If you are translating Figma designs into code, this saves serious time. Worth installing.

memcached-mcp-server9.8k★

Memcached MCP connects your AI assistant to Amazon ElastiCache Memcached clusters. Get, set, and inspect cache keys directly from your editor. Config is your cluster endpoint plus AWS credentials. ElastiCache bills per node-hour. The MCP is free. AWS Labs maintains it alongside the rest of their MCP suite. Useful for debugging cache issues without writing throwaway scripts. The catch: Memcached is simpler than Redis but also less capable. If you are already on Valkey or Redis, this is not the server you want.

amazon-keyspaces-mcp-server9.8k★

Connects your AI assistant to Amazon Keyspaces, AWS's managed Cassandra-compatible database. Create tables, run CQL queries, and manage schemas through natural language instead of wrestling with the Cassandra CLI. Setup needs AWS credentials and a Keyspaces endpoint. If you already use Keyspaces, this slots right in. The MCP server is free. Keyspaces pricing is pay-per-use based on reads, writes, and storage, with a modest free tier. Maintained by AWS Labs. Useful if you're already running Keyspaces workloads. For everyone else, there are simpler database MCP servers to start with.

aws-transform-mcp-server9.8k★

Connects your AI assistant to AWS Transform, the service that automates code modernization and migration tasks like Java upgrades and .NET porting. The MCP exposes operations for managing transformation workspaces, kicking off jobs, wiring up connectors, handling human-in-the-loop tasks, and downloading artifacts. Setup is the standard AWS Labs MCP pattern: install the package, configure AWS credentials, point your MCP client at it. Maintained by AWS Labs as part of the official AWS MCP suite. AWS Transform itself is metered by AWS. The free tier covers some monthly usage; beyond that you pay per transformation job, with separate pricing for HITL review hours. Check the AWS Transform pricing page for current rates. Install if you're already running AWS Transform jobs and want your AI assistant managing them from inside your IDE. Skip if you're not on AWS Transform; it doesn't do anything generic.

ecs-mcp-server9.8k★

Connects your AI assistant to Amazon ECS for container management. You can check service health, inspect task definitions, view running tasks, and troubleshoot deployment issues. Covers both Fargate and EC2 launch types. The MCP server is free and open source. ECS itself is free (you pay for underlying Fargate or EC2 compute). Setup requires AWS credentials with ECS permissions. Be cautious with write operations, stopping tasks or updating services in production needs careful IAM scoping. Maintained by AWS Labs. If you manage ECS services and spend time checking deployments or debugging task failures, this puts that workflow in your editor. The ability to inspect stopped task reasons conversationally is worth the setup alone.

eks-mcp-server9.8k★

Connects your AI assistant to Amazon EKS for Kubernetes cluster management. You can check cluster status, inspect node groups, review workloads, and troubleshoot pod issues. Bridges the gap between kubectl and your AI assistant. The MCP server is free and open source. EKS charges $0.10/hour per cluster, plus EC2 or Fargate compute costs. Setup requires AWS credentials, a configured kubeconfig, and the right RBAC permissions. More setup than most MCP servers. Maintained by AWS Labs. Kubernetes debugging is notoriously painful. Having an AI assistant that can inspect pod logs, check events, and correlate cluster state is valuable. The setup overhead is worth it if you're running EKS daily.

mobile-mcp9.1k★

MCP server for automating iOS and Android devices. Your AI assistant taps buttons, reads screen content, takes screenshots and navigates apps through native accessibility APIs rather than computer vision, which is why it is fast and does not break when a button moves two pixels. Needs Node.js 22+, plus Xcode for iOS or the Android SDK. One npm install, then point your MCP client at it. Works against emulators, simulators and real devices wired to your machine. Maintained by Mobile Next, a commercial company, not a community project, and the server itself is Apache 2.0. Free and the best option for driving mobile apps from an assistant. The catch is what sits next to it: Mobile Next also sells a paid cloud for real-device sessions billed per device-minute, and their Mobile CLI ships under the Functional Source License rather than Apache. The MCP server is genuinely open; the surrounding toolchain is not uniformly so.

codex-with-chatgpt7.4k★

Arbitrage on quota you already own. A ChatGPT Plus or Pro subscription has web usage sitting idle while your coding agent burns metered Codex tokens on planning and review. This bridge moves the thinking to the subscription and leaves execution with Codex. Your workspace is exposed over an OAuth-protected, read-only MCP connection. ChatGPT reads exactly the lines it asks for and hands back a plan; the repo is never uploaded and never writable. One-paste install, no API key, no reverse proxy. Maintained by an independent developer, not OpenAI, with bilingual Chinese and English docs. Install it if you run Codex daily and keep hitting planning limits. The catch is the trust boundary: a third-party bridge with live read access to your working tree is a call most security teams will make for you.

XcodeBuildMCP6.5k★

Lets your AI assistant build, test, and manage Xcode projects directly. Run builds, launch simulators, install on devices, and validate macros, all through MCP tool calls from Claude or Cursor. Install via Homebrew or npm, point it at your Xcode workspace, and it runs as a per-workspace daemon. Setup takes under five minutes. Official Sentry team maintains this. Free, open source. If you do iOS development with an AI assistant, this removes the constant terminal switching. Genuinely useful for the Xcode workflow.

devspace5.2k★

devspace is a self-hosted MCP server that gives an AI assistant like ChatGPT controlled access to your local projects. Instead of uploading your code anywhere, it exposes read, edit, search, and run operations on files already on your machine through a secure tunnel. MIT-licensed and free. What you get is a cloud model working directly against a real codebase on your disk: it can edit files, search the tree, and run commands in your project. Setup is moderate, you run the server locally and stand up the tunnel yourself, but once it is connected the assistant operates on your actual code rather than pasted snippets. It is a community project, not an official vendor integration, and the server itself is free. The cost is whatever AI client and model you point at it. This is worth it if you want ChatGPT or another MCP client working on local code without sending it to the cloud, and pointless if you already use a built-in agent like Claude Code that does this natively. The catch is the security model. You are handing an AI read, write, and execute access to your machine through a tunnel, so the convenience is only as safe as that setup. Trust the tunnel and the allowlist before you trust the workflow, and do not point it at anything you cannot afford to have an agent touch.

exa-mcp-server5.1k★

Connects your AI assistant to Exa's neural search API. Run semantic web searches, find similar pages, and get cleaned content from URLs. Think of it as giving your AI a research assistant that actually understands context. Setup is simple: get an Exa API key, add it to your config, and you're searching. The MCP server is free. Exa's API uses credit-based pricing, so your costs scale with usage. Maintained by Exa Labs (the company behind the API). Worth installing if you want your AI to pull real-time web data during conversations. One of the more practical MCP servers out there.

frontend4.6k★

Google's MCP frontend library for building client-side MCP integrations. Provides the tooling to connect applications to MCP servers, handle protocol negotiation, and manage tool discovery from the client side. This is a library, not a standalone server. You integrate it into your own frontend application to consume MCP servers. Setup complexity depends on your application architecture. Maintained by Google as part of their MCP ecosystem. Useful if you're building a custom AI interface that needs to talk to MCP servers. Most developers will use existing clients like Claude Code instead of building their own.

server4.6k★

Google's official MCP server for Google Cloud services. Connects your AI assistant to BigQuery, Cloud SQL, Spanner, Cloud Storage, and other GCP resources. Run queries, manage infrastructure, and inspect services through natural conversation. Requires Google Cloud credentials and appropriate IAM permissions. The MCP server is free. The underlying Google Cloud services have their own pricing, ranging from generous free tiers (BigQuery's 1TB/month free queries) to significant enterprise costs. Maintained by Google. Essential if your stack runs on GCP. The breadth of service coverage makes this one of the more ambitious cloud provider MCP servers.

csharp-sdk4.6k★

The official C# SDK for building MCP servers and clients. If you're in the .NET ecosystem and want to expose tools, resources, or prompts to AI assistants, this is the foundation you build on. Add the NuGet package, define your tools as methods, and the SDK handles protocol negotiation and transport. Clean API design that feels native to C#. No external service dependencies. Maintained by the Model Context Protocol organization, the same team behind the spec itself. Essential if you're building MCP integrations in C# or .NET. Skip it if you're consuming servers, not building them.

spec-workflow-mcp4.3k★

Brings spec-driven development into your AI workflow. Define requirements as structured specs, then let the MCP server guide implementation step by step. Turns vague feature requests into concrete, trackable development tasks. Setup is straightforward. Point it at your project, write specs in the expected format, and your AI assistant becomes a disciplined developer that follows the plan instead of freestyling. Community-maintained by Pimzino. Interesting concept for teams that want more structure in AI-assisted coding. Worth trying if you find AI assistants too eager to write code before understanding requirements.

excel-mcp-server4.2k★

Connects your AI assistant to Excel and spreadsheet files. Read, write, and manipulate .xlsx files through natural language. Create sheets, update cells, apply formulas, and extract data without opening Excel. Setup is minimal. Install the server, point it at your files, and start asking questions about your spreadsheets. Works with local files, no cloud service needed. Community-maintained. A practical pick for anyone who deals with spreadsheets regularly and wants AI to handle the tedious parts. Particularly useful for data extraction and report generation workflows.

vsix-tools3.8k★

Tools for packaging and managing VS Code extensions (.vsix files) through MCP. Build, validate, and inspect extension packages without memorizing the vsce CLI flags. This is a narrow-purpose server aimed at VS Code extension developers. Setup is straightforward if you're already in the extension development workflow. Maintained by Microsoft as part of their MCP monorepo. Only worth installing if you actively build and publish VS Code extensions. Everyone else can skip it.

vscode3.8k★

Microsoft's MCP integration for VS Code. Connects your AI assistant to the editor itself, enabling workspace management, file operations, and editor commands through the MCP protocol. Works within the VS Code ecosystem. Setup depends on your VS Code configuration, but the integration is designed to be seamless with Microsoft's AI tooling. Maintained by Microsoft. Relevant if you use VS Code as your primary editor and want deeper AI integration beyond what's built into Copilot. Part of Microsoft's broader MCP strategy.

spelling3.8k★

Runs cspell (Code Spell Checker) across your repository through MCP. Catches typos in code, comments, documentation, and configuration files. Uses the same dictionary infrastructure that powers the VS Code spell checker extension. Setup is minimal. The server uses existing cspell configuration if you have one, or works with sensible defaults. No external service needed. Maintained by Microsoft as part of their MCP monorepo. Handy for CI-style spell checking during AI-assisted code reviews. A small but useful addition to your MCP toolkit.

fastmcp3.3k★

FastMCP is a TypeScript framework for building MCP servers. Instead of wiring up JSON-RPC handlers and transport layers yourself, you define tools and resources with decorators and let FastMCP handle the protocol plumbing. Think Express.js but for MCP. Setup is npm install plus a few lines of code. You get type-safe tool definitions, automatic schema generation, and built-in validation. The API is clean enough that a working MCP server takes under 50 lines. Use this if you're building MCP servers in TypeScript. It cuts the boilerplate dramatically. The catch: you're adding a dependency layer between your code and the protocol, so breaking changes in FastMCP mean updating your servers.

supabase-mcp3.0k★

Supabase MCP connects your AI assistant to your Supabase project, covering database queries, auth management, storage buckets, and edge functions. One config file, one API key, and your assistant can run SQL, manage users, and check storage without you opening the dashboard. Community-maintained, not official Supabase. That said, it tracks the Supabase API closely and covers the core operations. Self-hosted Supabase instances work too. Setup takes about two minutes if you already have your project URL and service key. Install this if Supabase is your backend. Skips the dashboard for 80% of what you do. The catch: community project, so new Supabase features may lag behind.

x64dbg-mcp-server2.5k★

Connects a debugger rather than a SaaS API: x64dbg over MCP, so an assistant can set breakpoints, step, read memory, dump registers, scan patterns, walk xrefs, and inspect PE, PEB, and SEH structures. 84 tools, 22 event callbacks. Setup is a file copy. Drop the plugin into the x64dbg folder and it starts on launch with a bearer token generated on first run. Written in Zig as one native binary, so no .NET or Python runtime. Both x64dbg and the plugin are free, community maintained by a single developer. Worth installing for reverse engineering and malware analysis work. One caveat: the server binds 0.0.0.0 by default, so change that before running it anywhere but an isolated lab machine.

tavily-mcp2.4k★

Tavily's MCP server gives your AI assistant real-time web search, page extraction, site mapping, and crawling. Ask your assistant to research something and it pulls live results instead of relying on training data. Official, maintained by Tavily. Setup requires an API key from tavily.com. Free tier gives you 1,000 searches per month. Paid plans start at $100/mo for higher volume. The search quality is solid for research and fact-checking tasks. Install this if your AI workflow involves regular web research. The free tier is enough to evaluate it properly. The catch: API key required, and heavy usage means paying. If you only need occasional searches, a free alternative might be enough.

gortex1.9k★

Gortex indexes your repos into a graph of symbols, call chains, routes and cross-service contracts, then serves it to your coding agent so it queries the graph instead of reading whole files. It claims 257 languages, but only about 31 get full parsing; most of the rest are signature-only. It's one Go binary with no Docker or database: curl install, then gortex init per repo. No service costs; it runs locally under Apache 2.0, maintained by one main developer with real contributors. Worth installing, but the default hook mode blocks your agent's own Read and Grep on indexed files, so use --hook-mode=enrich if that bites.

reverify1.3k★

reverify plugs into Claude Code, Cursor, or any MCP client and hands your agent a reverse-engineering toolkit that fact-checks itself. Ask an AI to read a binary and it invents offsets, structs, and behavior with total confidence. reverify makes the model propose and the bytes decide: every claim comes back VERIFIED, REFUTED, or INCONCLUSIVE with the evidence it actually read. The core is pure Python, so PE/ELF/Mach-O parsing, disassembly, pattern scanning, and CPU emulation install with no Ghidra. Add the full extra and it upgrades itself in place to capstone, unicorn, lief, and Z3. It is free and MIT, and setup is a pip install plus an MCP entry. If you do malware analysis, CTF, or interop research and you have been burned by an AI confidently hallucinating a struct layout, this is worth wiring in. The catch: it grounds claims, it does not do the reversing for you. It is a lie detector, not an analyst.

platform1.2k★

The core platform component of Azure's MCP ecosystem. Manages Azure resource operations, subscription handling, and service discovery. Think of it as the backbone that connects your AI assistant to Azure's control plane. Requires Azure credentials with appropriate permissions. The MCP server is free. You pay for whatever Azure resources you provision or query through it. Maintained by Microsoft's Azure team. Essential if you're using Azure MCP servers, as other components depend on this platform layer. Not a standalone install, part of the broader Azure MCP project.

spelling1.2k★

Spell checking component within Azure's MCP monorepo. Runs cspell across the codebase to catch typos in code, docs, and configuration. Same approach as Microsoft's MCP spelling server, applied specifically to the Azure MCP project. This is a development tool for the Azure MCP project itself, not something end users install separately. It ensures code quality within the repo. Maintained by Microsoft's Azure team. You don't need this unless you're contributing to the Azure MCP project. It's internal tooling that happened to ship in the monorepo.

wrapper1.2k★

A wrapper layer in Azure's MCP ecosystem that provides standardized access patterns across Azure services. Handles authentication, request formatting, and response normalization so individual service integrations don't reinvent the wheel. This is infrastructure code, not a standalone server you install directly. It exists to support other Azure MCP components. Setup happens through the parent Azure MCP project. Maintained by Microsoft's Azure team. You won't interact with this directly. It's plumbing that makes the other Azure MCP servers work consistently.

git-mcp-server241★

This MCP server gives your AI assistant direct access to Git operations: commits, branches, diffs, logs, and file history. Your assistant can inspect repo state, compare branches, and read commit history without you running commands. Community-maintained and works with any Git repo on your local machine. No API keys, no remote service, just point it at a repo. Setup is one config entry with a path to your repository. Install this if you want your AI assistant to understand your repo's history and current state. Useful for code review workflows and understanding unfamiliar codebases. The catch: local repos only, and write operations (commits, merges) need careful permission scoping to avoid accidents.