MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
Local MCP server for accessible Parks Canada search and citizen-controlled booking preparation
npm MIT -
can-seegithub.com/HurleySk/can-seeupdated this year
MCP server that lets Claude Code see and interact with terminal apps
npm MIT -
can-seegithub.com/hurleysk/can-seeupdated this year
MCP server that gives AI agents eyes on your terminal — PNG screenshots, visual diffs, GIF recording, and full keyboard/text interaction
npm MIT -
Stream-first CAN & J1939 toolkit for security research and AI agents — typed JSONL events, CLI, TUI, fuzzing, MCP server.
pypi GPL-3.0 ★ 5 -
CLI-first CAN security research toolkit
pypi GPL-3.0-or-later -
Professional Serial Port Debugging Assistant with Modbus, PLC, CAN, I2C/SPI support and MCP server
NOASSERTION ★ 41 -
Free dangerous goods packaging API and MCP server by Fortitude21. No key, open CORS. UN number to permitted UN packaging (ADR 2025), UN mark decoder, ADR 1.1.3.6 points, IBC inspection dates, bund and spill pallet sizing, jerry can quantities.
MIT -
mcp_link_blendergithub.com/AuraFriday/mcp_link_blenderupdated this year
Let AI drive Blender using MCP - Your favorite LLM can now do anything Blender can do!
GPL-3.0 ★ 8 -
Long-term context and insights on your time-series data, such as telemetry, logs and sensor recordings: an MCP server that lets your AI agent query it all with SQL, on your machine. CSV, Parquet, JSON Lines, ROS 2, MCAP, PX4, ArduPilot, MAVLink, CAN with DBC, MDF4, NMEA 0183/2000.
Apache-2.0 ★ 2 -
Check your machine can actually run this project , before you waste a day finding out it can't. Zero-config dev environment preflight checker.
MIT ★ 2 -
Terminal UI engine for Go that tests can click and AI agents can drive (MCP). Zero-allocation rendering, immediate mode + Elm architecture, 3D, images, charts, accessibility, WebAssembly.
go Apache-2.0 ★ 195 -
A manager for code snippets and AI prompts that you can search by describing what they do. Claude Code and Cursor can search it too, over MCP. Next.js, Prisma, Postgres, Claude API.
MIT ★ 1 -
A video editor your coding agent can drive: a real timeline, frames it can look at, delivery checks, and a window where you watch and talk to it.
NOASSERTION -
Workflow-level MCP server for Vikunja: a gated Backlog→Queue→Design→Build→Review pipeline that AI coding agents can be left alone with. Agents can never move a task to Done.
MIT ★ 4 -
Your dashboard is data. Any AI can build it; any human can edit it. A protocol + runtime for agent-composable dashboards — layout-as-data, one guarded control plane, sandboxed agent-authored widgets, MCP server included.
npm MIT ★ 20 -
Point an agent at your homelab and it can nuke a VM. proxmox-mcp gates every write before the Proxmox API sees it, so an AI client can operate the cluster without owning it.
npm MIT ★ 11 -
Native macOS video editor that coding agents can drive: everything the UI does, Claude Code and Codex can do through the bashcut CLI or MCP server. Layered timeline, captions, voiceover, plugins.
MIT ★ 80 -
Claude can now actually touch your Google Ads account — not just talk about it in theory. It can read your data, spot the waste, and make the changes you approve.
pypi MIT ★ 4 -
The open-source Calendly alternative that AI agents can book. Keyless MCP server + llms.txt so any agent can check your availability and book you. Self-hosted, real Google Calendar, Stripe, embeds, REST API.
MIT ★ 2 -
Browsing MCP server gated by AgentLock. The agent can read any page; the page can't drive the agent. Content from the web is tagged untrusted at the source, and URLs, form values and destinations must trace to the user, never to page content.
pypi AGPL-3.0 -
A private decoupling layer between AI and any database or dataset. The connection stays on your side, the AI can't write SQL, and it can only read views you approve. Sometimes less is more.
npm MIT -
Vyapar is a aggregator/agentic commerce layer based model which makes merchant catalog AI transactable by any AI buyer in the world with one click. Any Merchant in the world can opt in, can be used by anyone, but powered by Razorpay.
-
An exoskeleton for your AI coding agent - it remembers everything, sees your whole codebase, can undo any change and can't fake a result. Fewer wasted tokens, fewer broken builds. Claude Code, Cursor, Codex & more.
pypi GPL-3.0 ★ 5 -
Your AI can already see the screen — it just can't measure it. uisight is an MCP server for web/responsive UIs: live mobile+desktop sessions, a measurement engine (contrast, touch targets, theme drift) that reports findings as text, and a panel you and yo
npm MIT -
Agent memory you can audit. Every lesson cites its source, and a bad memory can be traced to where it came from and removed with a receipt. Local-first and open source, shared across Claude Code, Cursor and Cline over MCP.
npm Apache-2.0 ★ 7 -
Your AI can already see the screen — it just can't measure it. Live mobile+desktop sessions, a measurement engine (contrast, touch targets, theme drift) reporting findings as text, and a panel you and your agent share. MCP server for web/responsive UIs.
npm MIT ★ 130 -
The browser tool that can stop and ask you. A 2FA code, a CAPTCHA, a choice only you can make: it asks on your screen, then carries on in the same signed-in Chrome. 40 tools, MCP clients that run local servers, MIT.
npm MIT ★ 64 -
154 Web3 MCP tools for Claude Code, custody-free. Plus an agent wallet prompt injection can't drain: per-tx, daily, and on-chain (ERC-4337) spending caps the model can't bypass. EVM + Solana + Tron DEX/DeFi, perps, prediction markets, contract deploy + audit.
MIT ★ 7 -
JigsawStack Translationno update data
With our real time AI Translation, you can translate any text, & images conversation, or document to any language while maintaining the context, tone, and meaning with AI enabling high accuracy and reliability. We support 180+ language pairs! Check out the list of languages we support here: https://jigsawstack.com/docs/additional-resources/languages You can get your jigsawstack key here: https://jigsawstack.com/dashboard/
smithery -
Renders what a branch changed to your architecture as one picture a human can confirm at a glance, and gives a coding agent boundary verdicts it can run on itself before asking for review. Hexagonal / ports-and-adapters, Kotlin. CLI · MCP · web UI.
MIT -
Browser MCP - the browser tool that can stop and ask you. A 2FA code, a CAPTCHA, a choice only you can make: it asks on your own screen, then carries on in the Chrome you're signed into. 40 tools, any MCP client, MIT.
npm MIT -
Screen recording your AI can read. Free, open source Mac app: press a hotkey, show the bug, say what's wrong. Luna turns your screen, voice and clicks into a small MP4 plus a prompt any agent can read: Claude, ChatGPT, Cursor, Codex. Works over MCP too.
MIT ★ 4 -
Agent memory you can actually read — it lives in your own markdown, not a vendor's database. Retrieval over the notes you already wrote, and a credential vault agents can use but never see. Self-hosted, MCP, one Go binary. Works on any folder: Obsidian vault, Logseq graph, or plain ~/notes.
go MIT ★ 3 -
Exegete: agentic qualitative data analysis application. You work in conversation with your AI assistant. Extensive human oversight routines and integrity checks are built-in. Files can be pseudonymised before the agent reads them. Works with cloud and local models. Projects remain compatible with QualCoder so that work can continue there as well.
pypi LGPL-3.0 ★ 7 -
Your agent can read git log. It can't read the shell commands you tried that failed. NexusMem indexes shell history with exit codes, git commits and diffs, and project docs into a local SQLite database, and serves back a token-budgeted slice on demand. No cloud, no account.
npm MIT ★ 14 -
MCP server exposing a curated, read-only Home Assistant entity allowlist to a low-trust client, via a guest secret that can't reach any other HA API. The MCP server can for example be used for a Chatbot available to guests of a Airbnb that is managed by Home Assistant.
Apache-2.0 ★ 1 -
MCP server for AI memory, governance, and continuity across session-death. A self-verifying chronicle: honest on write, on read, on reach, and about itself — a failed write can't report success, a partial answer can't pass as complete, and the capability surface is generated from the live registry, not typed. 100% local.
NOASSERTION ★ 3 -
On-chain spending firewall for AI agents on NEAR. The agent's money sits in a vault with no master key; its key can only ask to pay. A contract checks every payment (caps, allowlist, shared scam registry) and the rules can be locked for good. If the AI is fooled, the contract still says no.
npm MIT -
An open source Model Context Protocol (MCP) server and command-line tool that helps your AI coding assistants quickly create baseline IAM policies that you can refine as your application evolves, so you can build faster. IAM Policy Autopilot analyzes your application code locally to generate identity-based policies for application roles, enabling faster IAM policy creation and reducing access troubleshooting time. IAM Policy Autopilot supports applications built in Python, Go, TypeScript, and Java.
pypi Apache-2.0 -
MCP server that can execute terminal commands
NOASSERTION ★ 91 -
Recallgithub.com/adityapall1/Recallupdated this year
Recall is a personal knowledge base you can talk to. Save notes, snippets, and links, then search them by meaning rather than exact wording. Recall runs as an MCP server, so it works directly with your AI assistant, so you can ask Claude "what did I save about hash tables?" and get an answer from your own notes. There is also a CLI and web app.
MIT ★ 1 -
MCP Server that can generate Star Trek Klingon Lorem Ipsum
-
A mcp server which can get weather forcast and alerts
-
MCP Tools that can help on Java development
★ 2 -
imap-smtp-mcpgithub.com/mrworf/imap-smtp-mcpupdated this year
MCP Server that can connect to IMAP and SMTP
GPL-3.0 -
iflow-mcp_mcp-local-raggithub.com/nkapila6/mcp-local-ragupdated this year
MCP "server" that can perform a web search locally without the use of APIs.
pypi MIT -
Calibre MCP server that can interact with your Calibre Library via the Content Server
★ 2 -
Provide MCP services via API, can be used for the Joplin plugin NoteLLM.
★ 2 -
claude-gmail-mcpgithub.com/pliablepixels/claude-gmail-mcpupdated this year
Super simple MCP gmail server for claude, so it can send messages
MIT -
Open Banking MCP server I made so I can chat with/plan my finances easier.
MIT ★ 1 -
Make your agent can search and read web without overhead
MIT ★ 1 -
undo-mcpgithub.com/craig-whitfield/undo-mcpupdated this year
MCP server for reversibility intelligence — check if actions can be undone
npm -
deepwikigithub.com/a-khushal/deepwikiupdated this year
Paste a GitHub URL. Get a wiki. Ask questions it can actually answer.
-
tarot_mcp_servergithub.com/hwfish/tarot_mcp_serverupdated this year
A MCP tool that can provide LLM randomized tarot draws and sufficient description.
-
MCP file server so Claude Cowork can read and write ~/dev in WSL
MIT -
MCP server that joins a live Excalidraw collaboration room so an AI agent can read and write the drawing
npm MIT -
An MCP server that can only read and write files inside allowed folders
pypi MIT -
MCP server that can execute commands such as keyboard input and mouse movement on macOS
MIT ★ 61 -
This is a MCP service that can be remotely called, built on FastMCP.
★ 1 -
A collection of purchasing agents that can help with customer enquiries.
GPL-3.0 ★ 1