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.
-
MCP Kanban is a specialized middleware designed to facilitate interaction between Large Language Models (LLMs) and Planka, a Kanban board application. It serves as an intermediary layer that provides LLMs with a simplified and enhanced API to interact with Planka's task management system.
★ 65 -
Local LLM chat history, served over MCP
Apache-2.0 ★ 2 -
Your personal LLM wiki: a git-versioned Markdown vault served over MCP
AGPL-3.0 ★ 1 -
Vector illustration engine for LLM agents, served over MCP.
npm MIT -
Memory database for LLM agents — persistent semantic + keyword memory, project namespacing, served over HTTP REST and MCP.
pypi AGPL-3.0-only -
Zero-LLM structural code retrieval for AI coding agents, served over MCP
pypi MIT -
Crawl, index, and hybrid-search documentation sites; served to LLM agents as an MCP server. FastAPI + DuckDB + BM25/embeddings, fully offline-testable.
MIT -
Benchmark local LLM inference speed (tokens/sec) on your own hardware — llama.cpp native + cloud APIs, 124-model catalog, optimal-quant picker, and an MCP serve mode.
MIT ★ 2 -
GitHub-Documentation-Rag-Agentgithub.com/Mandoa-Labs/GitHub-Documentation-Rag-Agentupdated this year
RAG agent that ingests GitHub's official documentation, embeds it with local Ollama LLMs, and serves semantic search via a PostgreSQL pgvector database.
-
Explainable, vectorless RAG for AI agents: BM25 + temporal knowledge graph over your files, served over MCP. R@5 0.977 on LongMemEval-S, ~2 ms queries, 0 LLM calls, zero dependencies.
MIT ★ 1 -
One binary that reads offline Wikipedia, and any other ZIM archive, out of a local Kiwix server. kiwix-agent mcp speaks MCP over stdio to any client; kiwix-agent chat serves a browser chat page driven by a local LLM.
MIT ★ 2 -
Turn a codebase into a deterministic AST knowledge graph (no embeddings, no LLM) and serve it over MCP, so coding agents ask the graph instead of grepping.
MIT -
jupyter-interpreter-mcpgithub.com/lmseidler/jupyter-interpreter-mcpupdated this year
Serving access to a JupyterHub as code interpreter sandbox for LLMs using MCP.
MIT -
Memory database for LLM agents — persistent keyword + optional semantic memory, project namespacing, served over HTTP REST and MCP from a single ASGI process. SQLite-backed, packaged as a Docker container. Runs on your hardware.
AGPL-3.0 ★ 16 -
Agent memory that needs no API key, no LLM, and no embedding provider. One command on Bun and it is serving MCP over stdio against a local SQLite store. Drop-in for the MCP reference memory server. Every memory keeps its source, who may read it, and the evidence that contradicts it.
npm Apache-2.0 ★ 16 -
Stratagithub.com/avinash-1707/Strataupdated this year
Local-first memory layer for AI agents. Compresses context with a local LLM, indexes it with hybrid search, and serves it over MCP.
MIT -
CodeRadargithub.com/officialCodeWork/CodeRadarupdated this year
React & TypeScript static analysis that maps every UI component to the APIs, state, and events behind it — and serves that context to AI coding agents over MCP. Screenshot or ticket in; component, data lineage, user journeys, and blast radius out. Deterministic, no LLM in the core.
MIT -
Zero-LLM structural code retrieval for AI coding agents — tree-sitter index, BM25+dense+structural fusion, served over MCP.
MIT ★ 7 -
Agentic control layer for the Censurado self-hosted AI news portal: a stdlib Python CLI plus an MCP server (32 tools) an LLM agent drives to write and publish AI-persona articles through a gated editorial walk, curate the front page and authors, and deploy. Ships the prompt/skill recipe and a 24/7 serve loop.
MIT ★ 1 -
@wall.cx/cliactive
Wall data reader + self-hosted MCP server (single binary). Reads Wall export ZIPs and serves them locally for personal viewing, search, and LLM access.
npm MIT -
UK Statutory Compliance Almanac & Property Tribunal RRO Precedent Finder on Cloudflare Workers — serving /rro, /llms.txt, /.well-known/agent.json, and deterministic statutory audit APIs.
-
mathlas-mcpgithub.com/archerkattri/mathlasupdated this year
A tool FOR an AI (no API key, no LLM): search existing math over a 3.7M-doc index + airtight numeric/Lean verification + mathlib search (Loogle/LeanSearch) + OEIS/PSLQ identification + needs<->guarantees scaffolds, served over MCP.
pypi Apache Software License -
Agent memory with no API key, no LLM, and no embedding provider. Serves MCP over stdio against a local SQLite store, or Cloudflare Workers + D1. Drop-in for @modelcontextprotocol/server-memory; every memory keeps its source, its scope, and the evidence th
npm Apache-2.0 -
agentedgithub.com/frane/agentedno update data
A text editor for LLMs, not humans. Take ed, the line editor that nobody has voluntarily used since about 1975, and rebuild it for an environment where the typing user is a language model. Short verbs, line addresses, no modes, no TUI. What an editor optimises for changes when the user is the model: round trips per task, tokens per command, an editing buffer with a long memory, and an undo tree that remembers the branches the agent abandoned, because that's often where the interesting work was. ## Features - **Fewer round trips.** Read-before-Edit is unnecessary; on conflict the response carries the new content so you reconcile in one call instead of pre-reading every time. - **Branching undo.** Walked-back work stays addressable instead of being thrown away when you pick a different path. - **Three-way merge.** Concurrent agents get a structured conflict response instead of a silent overwrite. - **Atomic batches.** Multi-file refactors run all-or-nothing instead of leaving half-applied state on failure. - **Cross-file moves and regex replace as primitives.** Operations the built-in tools can't express cleanly become single calls. - **Drift detection.** External edits to an open file are folded into the tree instead of being clobbered by the next write. - **Inline diagnostics.** Type errors and lint findings surface on save, not at the next build many edits later. - **Cross-session memory.** Per-file notes persist between sessions and surface inline on the next open. - **Audit log.** Every operation recorded with actor and timestamp, so two agents in one workspace can't argue about who moved the head. ## Multi-workspace (v0.4.0) One MCP server registration handles every project you mention. Tool calls route to the workspace owning the absolute path you pass; each project gets its own SQLite, its own undo tree, its own LSP daemon — lazily spawned on first touch. Desktop hosts (Claude Desktop, Codex Desktop, Cursor) install the server once globally and use it across projects without per-project config. ## Install `ae` must be installed locally; this MCP bundle registers `ae serve` as a stdio server. ```sh brew tap frane/tap && brew install agented # or curl -sSL https://raw.githubusercontent.com/frane/agented/master/install.sh | sh ``` Pure Go, single static binary, Apache 2.0. Source: https://github.com/frane/agented
smithery -
Badgr — GPU Compute, AI Model Deployment & Agent Executionno update data
**Run GPU workloads, deploy AI models, access inference APIs, and execute AI agents using Badgr's MCP server.** Badgr is an AI compute and workload execution platform for developers, coding agents, and AI assistants. Run workloads on cloud GPUs or compatible customer-owned GPUs (BYO GPU), with cost controls, execution logs, and verified results. **What can Badgr do?** - **GPU Compute / AI Workloads:** Run Python scripts, machine learning jobs, batch inference, and supported ComfyUI workflows. - **AI Model Deployment:** Deploy and serve LLMs using compatible runtimes such as vLLM and SGLang. Get a verified, OpenAI-compatible inference endpoint. - **AI API / LLM Inference:** Make AI API requests with response validation, supported fallback routing, and bounded execution. - **AI Agent Execution:** Launch coding and testing agents, run tasks, check acceptance criteria, and retrieve execution receipts. **Key features:** Cloud GPU execution, BYO GPU support, model serving, AI agent jobs, workload verification, execution logs, cost limits, and deployment management. Connect Badgr to your MCP-compatible AI assistant to execute real workloads instead of just generating instructions. Get started: https://aibadgr.com **Requires a Badgr API key. GPU workloads and other paid execution may incur usage charges.**
smithery -
Local code index for AI agents. Scans a repo with tree-sitter into one SQLite file — function inventory, call graph, module deps, control flow, impact analysis — then serves 32 MCP tools to Claude Code, Cursor and VS Code. No LLM calls, no upload, no telemetry. npm i -g codindex
MIT ★ 3 -
mcp_command_servgithub.com/jon2allen/mcp_command_servupdated this year
python based MCP cli tool
MIT ★ 1 -
AI-Inference-Managed-by-AI: Go binary for managing AI inference on edge devices
Apache-2.0 ★ 19 -
DIOgithub.com/nisaral/DIOactive
Drop-in OpenAI- and Ollama-compatible LLM gateway that learns each backend's latency online and routes vLLM / SGLang / TGI / Ollama with SLO-aware admission. No engine patches.
Apache-2.0 ★ 3 -
sushimcpgithub.com/maverickg59/sushimcpupdated this year
SushiMCP is a dev tools MCP that serves context on a roll.
AGPL-3.0 ★ 9 -
SushiMCP a dev tools model context protocol server that serves context on a roll.
npm AGPL-3.0-or-later -
ragkitgithub.com/cognis-digital/ragkitupdated this year
Batteries-included local RAG pipeline — ingest, index, serve
NOASSERTION -
Model Context Protocol server that serves Django documentation to AI coding agents
npm MIT -
Model Context Protocol server that serves Livewire documentation to AI coding agents
npm MIT -
Model Context Protocol server that serves Filament v5 documentation to AI coding agents
npm MIT -
A cringe-worthy MCP server that serves ads to developers in Cursor, Claude, and other clients
npm MIT -
MCP server serving NVIDIA CUDA-Q documentation, API reference, and runnable examples to AI agents, version-pinned to your installed cudaq.
pypi Apache-2.0 -
Self-hosted, read-only MCP server that serves IMAP mail from a local notmuch index over authenticated HTTP
MIT ★ 20 -
Invite an AI agent into a Pyunto diary: replies in the thread or serves Claude via MCP. Decrypts only on your machine.
Apache-2.0 ★ 1 -
An email client for AI agents. Filtered, read-only Gmail delivered to disk as .eml + markdown, and served over MCP.
MIT ★ 3 -
Fast, low-token web reader and focused crawler for AI assistants, served over MCP. Returns ranked passages instead of whole pages.
Apache-2.0 ★ 7 -
EKOS is an AI-native platform that continuously reconstructs, compiles, stores and serves enterprise knowledge.
MIT ★ 28 -
akopiagithub.com/rlabs-cl/akopiaupdated this year
Plug. Sync. Serve. — headless retrieval plane for your AI stack. Self-hosted RAG infra with Qdrant + Meilisearch + Redis, plugin-first, MCP-native.
MIT -
Reliable verified commands for Claude Code. Deterministic verify + retry + fallback over hermetic claude calls, served as MCP tools.
npm MIT -
Remote (HTTP) MCP server for Signal Messenger — serve signal-mcp to claude.ai, mobile and other remote clients
MIT -
agrosagegithub.com/prashant-cr/agrosageupdated this year
Multi-agent crop advisory for smallholder farmers: weather + mandi prices + agronomy rules, orchestrated by Claude or served over MCP. Zero-key offline demo.
MIT ★ 1 -
Private, asynchronous mailbox for AI agents. Reference implementation of SAMP (Secure Agent Messaging Protocol), served as an MCP server on Cloudflare Workers.
AGPL-3.0 -
A local MCP server that serves your AI only the rules matching the file you're editing — just-in-time, per file.
MIT -
architecture-intelligence-platformgithub.com/michaelegner/architecture-intelligence-platformactive
Trusted architecture context for AI agents: an evidence-backed architecture graph from OpenAPI, AsyncAPI and OpenTelemetry, served over MCP.
Apache-2.0 ★ 10 -
The web toolkit for AI coding agents — scrape, search, map, crawl, index, extract, diff, serve. Single static binary, MIT. Self-hosted Firecrawl/Exa/Tavily alternative.
MIT -
@mashware/domain-memory-servergithub.com/mashware/domain-memoryupdated this year
Domain Memory MCP server (stdio). Local-first MCP server that serves the business-domain knowledge of a project to any MCP-compatible agent.
npm Apache-2.0 -
Reduce MCP tool overload with a single Go binary that indexes tools locally and serves search-before-invoke via SQLite
MIT -
MCP server for GeneReviews: the expert-authored, peer-reviewed gene–disease chapters on NCBI Bookshelf, served as a searchable corpus of individually citable passages.
MIT -
CodeCortexgithub.com/rushikeshmore/CodeCortexupdated this year
Persistent codebase knowledge layer for AI agents. Pre-builds architecture, dependency, coupling, and risk knowledge served via MCP. 27 languages, 13 tools.
MIT ★ 10 -
BETA — Local-first memory for AI agents. Serves records with their source, date, and whether they actually answer the question. No account, no API key, nothing leaves your machine.
BSD-3-Clause -
In-memory document store served over Model Context Protocol (stdio). Go rewrite of the Python mcp_server from Anthropic's MCP introduction, plus a completion handler, resource templates, and multiple prompts.
MIT -
localgraphgithub.com/mehuljn/localgraphupdated this year
LocalGraph is a zero-dependency CLI tool written in Go that parses repositories into SQLite relationships (Imports, Definitions, Calls). It computes semantic embeddings locally using LM Studio or Ollama and serves codebase context directly to AI editors (Cursor, OpenCode) over the Model Context Protocol (MCP).
-
Local video tools for AI agents that read text and images but not video: contact sheets, transcripts, cut, loudness and sync checks, served over MCP.
pypi MIT ★ 1 -
liengithub.com/getlien/lienactive
Local-first semantic code search for AI coding assistants. Indexes your codebase with local embeddings and serves context over MCP — find code by meaning, check impact before editing, and surface complexity hotspots, all without sending code externally.
AGPL-3.0 ★ 2 -
Git-like versioning for prompts with eval-gated promotion. Store, diff, tag (draft/staging/production), A/B serve, and gate promotion on eval results. Self-hosted, not SaaS.
MIT