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.
-
Bilingual (English-Arabic) realtime voice AI agent for LiveKit using Azure Speech STT/TTS, GPT-4o-mini, and a FastMCP HTTP server for tool calling.
★ 1 -
Open infrastructure for Argentina's sociedades de IA (AI-operated companies, draft General Companies Law art. 14). 36 npm packages, 235 typed tools for the Vercel AI SDK, 6 RFCs, forensic audit log. MIT.
npm MIT ★ 3 -
tether-name-mcp-servergithub.com/tether-name/tether-name-mcp-serverupdated this year
catch the drift: verify AI agents are who they say they are. MCP server for tether.name
npm MIT -
arena-mcpgithub.com/ertekinno/arena-mcpupdated this year
Model Context Protocol server for Are.na - enabling AI agents to browse and interact with Are.na channels and blocks
MIT ★ 14 -
Shared memory for AI agents: Markdown documents in your own Supabase project are the source of truth; vector chunks are a rebuildable derived index. Agents connect over MCP, writes land in an inbox with provenance first.
npm MIT -
Local-first persistent memory for AI coding agents. Citations are verified at write time and stale memories are withheld at recall. Zero daemons, no API key, MIT.
npm MIT -
safe-solana-mcpgithub.com/liander-ai/safe-solana-mcpupdated this year
An MCP server giving AI agents safe, policy-gated access to Solana (with a Chain Fusion read on ICP): reads are open, transfers are capped, allowlisted, simulated, and returned unsigned. No keys, no broadcast.
MIT -
Read-only MCP server for VR.org. Gives AI agents one-call access to live VR, AR, and XR news, original editorial, headset deals, buyer guides, and top-game and top-app lists. Free, no keys, no writes.
npm MIT -
Lighthouse for AI agents — test if APIs are properly agent-accessible
pypi MIT -
Agents are in charge on MyMCPSpace.com
npm MIT -
Agents are in charge on MyMCPSpace.com
npm MIT -
dibsgithub.com/Agenxy/dibsactive
Tells agents what the others are working on, and where they collide.
go GPL-3.0 ★ 2 -
A macOS text editor where AI agents are first-class users. The agent edits, you decide.
MIT ★ 1 -
Open-source helpdesk where AI agents are the support staff and often the customers. Tickets, MCP, A2A, email, webhooks.
MIT -
An agent memory server that learns from feedback which memories are worth returning. Go, with Python/LangChain/MCP adapters.
MIT -
lobsterrollgithub.com/onEnterFrame/lobsterrollupdated this year
🦞 Agent-native messaging — where AI agents and humans are equal participants. Open source, self-hostable, MCP-ready.
npm Apache-2.0 ★ 10 -
Dependency-aware ticket board for humans and AI agents. Tickets are JSON files in your repo; agents claim work over MCP.
MIT ★ 1 -
Open-source issue tracker where AI agents are members of your team. Self-hostable, MCP built in.
AGPL-3.0 ★ 2 -
Open-source, DOM-native design canvas driven by agents over MCP. Designs are JSON files you own.
npm Apache-2.0 -
arca-agent-skillsactive
Agentic AI prompt instructions and MCP server for ARCA compliance via KeyCAE.ar — 12 tools for coding agents
npm MIT -
raysensegithub.com/RayforceDB/raysenseupdated this year
A structural X-ray for the codebases AI agents are writing.
cargo MIT ★ 14 -
Risk oracle for x402 endpoints — tells AI agents which paid endpoints are alive, honestly priced and safe to call. Zero runtime deps.
npm ★ 1 -
The shared brain your agents are missing — open-source context engine for AI agents
pypi Apache-2.0 ★ 1 -
Data Agents are intelligent assistants built by data engineers to help non-data professionals navigate the organization’s data infrastructu
MIT ★ 25 -
CRYPTYX for AI agents: hosted MCP server and skills for Claude Code, Grok Build and other agent hosts. Digital-asset intelligence; only coin symbols are shared.
MIT -
Corkboard of agentic post-it notes. A place for your AI agent to scribble bite-sized notes — a memory, so things it learns in one session are still around in the next.
MIT -
Open-source team chat where agents are members, not widgets. Self-hosted, Slack-familiar, MIT.
MIT -
@skr0l/mcpupdated this year
Official MCP server for SKR0L — the social platform where AI agents are first-class citizens. Post, browse, comment, and follow programmatically.
npm MIT -
Argus MCP server: lets AI agents check whether wallets are independent operators and how much ERC-8004 reputation is real.
npm MIT -
A browser-control layer that lets your AI agent drive the Chrome you are already logged into.
npm NOASSERTION ★ 8 -
shot-canvasactive
Shot Canvas on your computer: a design canvas you and your agent share, over MCP. No account - canvases are files here until you share them.
npm UNLICENSED -
Public message board for AI agents. One GET publishes. Humans read, agents write; the code, the moderation log and the dataset are all open.
Apache-2.0 ★ 2 -
MCP server exposing Omitly's local, verifiable PDF redaction to AI agents. Runs on-device — documents are never uploaded.
npm NOASSERTION -
Governed conversational analytics: dbt semantic layer, lineage + knowledge graphs, and an MCP server that lets an agent answer questions it can prove are right.
MIT -
Use a secret without holding it: an encrypted vault whose values are injected into the commands an agent runs and redacted from everything that comes back.
npm MIT -
brusegithub.com/padmk/bruseactive
Let local AI agents use your real browser in background tabs, where you are signed in. MCP server and Chromium extension.
MIT ★ 1 -
Remote MCP server: bring in a human expert (math, optimization, AI) when you or your agent are stuck. 20 USD per request.
-
A small self-hosted kanban board with an MCP server, so you can see what your coding agents are doing.
npm MIT ★ 1 -
MCP server for BitGraph: make and check BitGraphs from an AI agent. Files are hashed locally; only the SHA-256 is sent.
npm MIT -
MCP server: split inbound email threads and validate AR agent output
pypi MIT -
A small MCP server can assist AI agents in working with large GraphQL schemas that are retrieved through introspection.
GPL-3.0 -
Company AI assistant with a managed memory layer, permissioned retrieval, and agents that cite only what they are allowed to see.
MIT -
@golproductions/check-mcpgithub.com/golproductions/check-mcpupdated this year
Less hallucinations. A command firewall for AI agents: validates every shell command before execution, so fabricated commands are denied before they run.
npm SEE LICENSE IN LICENSE -
Validate AGENTS.md and CLAUDE.md against repo reality: referenced paths, package.json scripts, and documented commands are checked against the actual tree, so agent instruction files cannot quietly rot.
npm MIT -
Certify AI agents are safe for your internal web app before granting production access. CLI + MCP server, npm/PyPI.
pypi Apache-2.0 -
Content-addressed contracts with a hash-keyed, team-shareable verification cache for spec-driven agent loops, over MCP. Contracts are warp; code is weft.
pypi Apache-2.0 ★ 2 -
Official Lorde Studios Browser-use desktop browser and agent workstation for Windows. Automation, MCP server and agent chat. Closed source: builds and documentation are published here.
★ 1 -
3D & AR SDK for Jetpack Compose (Filament, ARCore), Compose Multiplatform, SwiftUI (RealityKit, ARKit) and the web, with Flutter and React Native bridges. The maintained successor to Google Sceneform. Opens glTF/GLB and USDZ, at real size in AR. Assistant-ready: MCP server, AGENTS.md and llms.txt.
npm Apache-2.0 ★ 1,374 -
Attach to the Chrome you are already logged into: a dependency-free CDP CLI and an MCP bridge for AI agents.
npm MIT -
agent-skills-mcpgithub.com/sohutv/agent-skills-mcpupdated this year
This project implements a Model Context Protocol (MCP) server that exposes "Skills" to AI agents. Skills are specialized instruction sets and resources that help agents perform specific tasks with domain expertise. You can find detailed skills specification here: [Agent Skills Spec](https://agentskills.io/specification)
Apache-2.0 ★ 11 -
MCP server for encrypted, persistent agent memory: facts are embedded and encrypted on the client, so the server stores only ciphertext and vectors.
pypi MIT -
Live experimental discovery run (ATSP World 0.1.0) to ascertain agent activity and its bounds: how autonomous agents behave and where their limits are.
pypi BUSL-1.1 -
MCP server that lets AI agents scan code, staged diffs and text for secrets with gitleaks. Findings are always redacted.
MIT -
MCP server for Bridle: lets an AI agent read its spending rules, check a payment, and propose one. Owner actions are deliberately not exposed.
npm MIT -
One CLI and MCP server for coding agents to find vulnerabilities across code, dependencies, apps, and infrastructure — on targets you own or are authorized to test.
npm MIT -
Pick the night your fans are free: concert dates ranked by Qloo audience affinity, with a Gemini agent and MCP server
MIT -
simple mcp server using python hosted in render used by n8n agent to generate staff roster in are workplace
-
A self-hosted, multi-tenant CMS built for AI agents: a website is data, not code. Sites are built over MCP, reviewed and published by humans.
BSD-2-Clause ★ 2 -
lifeline-mcpgithub.com/pattom/lifeline-mcpupdated this year
MCP server for on-demand human-in-the-loop interactions via Slack - get human input when AI agents are stuck or facing critical decisions
npm MIT -
MCP server for ShellMail — full email client for AI agents: read inbox, extract OTPs, search, send/reply, and permanently delete messages (destructive tools are annotated and require confirmation)
npm MIT