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.
-
—
MIT -
AI model router for agents. Checks whether a local model can handle the task before calling cloud inference. LOCAL/CLOUD verdict saves cost on every call.
npm MIT -
local-model-workers-mcpgithub.com/gaabrielrd/local-model-workers-mcpupdated this year
Servidor MCP que delega exploração de código, busca semântica, testes e correções a modelos locais, retornando resultados estruturados e diffs validados.
MIT ★ 1 -
Run, fit, and route work to local LLMs: one MCP server for Ollama, LM Studio, llama.cpp, Docker Model Runner, MLX, vLLM, and Foundry Local, with hardware-aware model selection.
npm CC-BY-NC-ND-4.0 -
MCP server for delegating mechanical tasks to local LLMs via Ollama. Claude does the thinking, your local model does the grunt work.
MIT ★ 15 -
Secure Model Context Protocol (MCP) server & gateway for Claude, Cursor, and ChatGPT. Unlimited local file editing, testing, and dev tools without cloud limits.
Apache-2.0 ★ 4 -
Zero-dependency local translation engine via Transformers.js NLLB model — translate text across 200+ languages, CPU-only, no cloud APIs, everything stays on your machine.
npm MIT -
A local model engine with a search engine and web crawler built in. Run your coding agents on your own GPUs, and get cited answers from your files, code, and the sites you crawl. One executable.
pypi MIT ★ 71 -
KnowledgeMCPgithub.com/maxzrff/KnowledgeMCPupdated this year
Model Context Protocol (MCP) server for local vector storage & semantic search (ChromaDB, OCR, async ingestion).
MIT ★ 8 -
Contract-driven coding workflow: the strong model plans and accepts, your local LLM does the coding. 25 MCP tools + 6 skills, 8 agent hosts (ZCode, Claude Code, Codex, VS Code, DSH, pi, omp, opencode). Local-first, no telemetry.
MIT -
A sleek C# class library designed to integrate a local Model Context Protocol (MCP) server with a local Large Language Model (LLM), tailored for Godot-based games. This library empowers your LLM to seamlessly interact with the game world through structured tool commands, bringing your Echoes of Ir-style vision to life.
MIT ★ 1 -
The home port for your local AI. Zero-install Mac app + MCP server: talk to Claude to discover, download and verify Hugging Face models straight to your own drive. By CAIO Group.
MIT ★ 2 -
Phocinae MCP decision server: gate / classify / route / score tools backed by the local Phocinae model
Apache-2.0 -
Google-Tasks-Local-MCP-Servergithub.com/brandon-lane-xyz/Google-Tasks-Local-MCP-Serverupdated this year
Local Model Context Protocol (MCP) server for Google Tasks
★ 1 -
A Model Context Protocol server for reading local files
npm MIT -
Give OpenAI agents the ability to securely spawn and govern local models like Qwen and DeepSeek
Apache-2.0 -
The local intern for Claude Code — 44 job-shaped tools, evidence-first briefs, durable artifacts. Local-first MCP server over Ollama with zero egress by default; no GPU big enough? Ollama Cloud runs the same 44 tools on 600B-class models, with automatic local fallback. Validated Hermes Agent integration on hermes3:8b.
npm MIT ★ 2 -
A lightweight, extensible Model Context Protocol (MCP) server for secure key management in local AI agents
MIT ★ 1 -
Offline terminal, browser, and web-search tools for any local LM Studio model — no cloud API, no API keys.
MIT -
Runs CodeZaiku, a coding and development harness on local models (Ollama, llama.cpp) or hosted APIs, as an MCP server, so Claude Code, Codex or Gemini CLI can hand it coding, review and ops tasks.
npm Apache-2.0 -
Local-first AI memory your assistants share. One store, every model, your machine.
NOASSERTION -
Local-first, cognitive persistent memory & knowledge graph for AI coding agents via Model Context Protocol (MCP)
AGPL-3.0 -
Remote MCP server for readygpu: live GPU, Mac and mini PC prices for running open AI models locally.
MIT -
Local-first, MCP-native tool for modelling software architecture. C4-style semantic model, React Flow editor, SQLite, single container — a self-hosted Structurizr alternative your AI client can edit over MCP.
MIT ★ 29 -
local-repl-mcpgithub.com/angrysky56/local-repl-mcpupdated this year
A locally-running Python REPL server that integrates with Claude Desktop through the Model Context Protocol (MCP). This runs entirely on your local machine with no need for API keys or remote execution.
MIT ★ 4 -
Deterministic, local-first context optimization for LLM prompts: clean, dedupe, structure, and measure token and cost savings before calling a large model.
pypi MIT -
TokenMark MCP server: expose local-LLM benchmark configs + hardware-aware model recommendations to Claude and other agents.
npm MIT -
A high-performance Model Context Protocol (MCP) server providing local speech-to-text transcription using whisper.cpp, optimized for Apple Silicon.
MIT ★ 20 -
gguf-mcpgithub.com/arose26/gguf-mcpupdated this year
MCP server that inspects local GGUF and safetensors model files — architecture, quantization, parameters, tensors, and VRAM-fit estimates from headers only, without loading the model
npm MIT -
MCP server wrapping local Ollama models - offload bulk generation, summarisation, classification, and code tasks from API-priced orchestrators to a model running on the user's own hardware.
npm Apache-2.0 -
axmggithub.com/dchrnv/axmgactive
Deterministic Merkle DAG associative memory engine with Model Context Protocol (MCP) for local LLMs and autonomous agents.
AGPL-3.0 -
TokenMark MCP server and CLI: measured local-LLM speeds and model picks for Strix Halo, DGX Spark and Mac
MIT -
aether-agentgithub.com/takzen/aether-agentupdated this year
A autonomous, local-first personal intelligence layer. Features an Active World Model, Qdrant Graph Memory, Circadian Rhythms, and MCP integration to act as the hidden brain for your digital life and IDEs.
Apache-2.0 ★ 9 -
Local-first workspace token calculator and model recommender for coding agents, with repo maps, VS Code, CLI, and MCP.
MIT ★ 46 -
A research prototype for a model-independent, verifiable shared knowledge layer for local or remote LLMs
Apache-2.0 -
KaiBoard MCP — Model Context Protocol server + command core, letting AI agents draw on local-first KaiBoard whiteboards.
MIT -
MCP server that gives a local model persistent memory: remember, recall, update and forget short facts stored in SQLite, launched by LM Studio
-
vlfs-localgithub.com/fedevela/vlfs-localupdated this year
A local, lightweight Model Context Protocol (MCP) server implementing the OpenViking standard. Turns any workspace into an intelligent, semantic virtual filesystem (VLFS) for LLM agents to prevent context window exhaustion.
-
Local offline speech recognition for Russian speech with German words: ASR model ensemble with word-level voting, diarization, desktop app, CLI and MCP server. Research notes included.
GPL-3.0 -
MCP server that delegates Claude Code subagents to local models (LiteLLM/llama.cpp/Ollama), DeepSeek, AWS Bedrock, or any OpenAI/Anthropic-compatible backend — without losing the Claude Code orchestrator session.
MIT ★ 19 -
Local-first SQL client for SQLite and PostgreSQL where AI agents work through guardrails: every write is previewed, approved in the app, logged and undoable. Built-in MCP server, Text-to-SQL with local models, schema visualizer. Tauri + Rust + React.
npm MIT ★ 8 -
Local data workbench a local language model drives over MCP: ingest, versioned cleaning, quality rules, charts, dashboards as code and quick models. Part of the Hoard family.
MIT -
local-mcp-crmgithub.com/bhairavaa/local-mcp-crmupdated this year
Local-first CRM built on the Model Context Protocol (MCP) — customer/project management exposed as MCP tools, usable from a custom LlamaIndex agent or directly in Claude Code or any other applications supporting MCP Servers or resources.
★ 1 -
Evidence-first local context engine for AI agents - durable project memory for code, documents, and model-adaptive ContextPacks.
NOASSERTION -
comfyui-mcp-servergithub.com/joenorton/comfyui-mcp-serverupdated this year
lightweight Python-based MCP (Model Context Protocol) server for local ComfyUI
Apache-2.0 ★ 412 -
An Open Sourced Model Context Protocol (MCP) Local server that gives AI Clients real-time cybersecurity reconnaissance capabilities
MIT ★ 29 -
my-localmcp: a local, offline Model Context Protocol (MCP) server that gives AI coding agents fast, deterministic repository context. SQLite indexing with optional Ollama re-ranking.
MIT ★ 1 -
mcp-local-school-orchestratorgithub.com/NxtGenCodeBase/mcp-local-school-orchestratorupdated this year
A Local Agentic School Management System using Model Context Protocol (MCP), LangGraph, and Ollama. Feature-rich PoC with SQLite data, PDF reporting, and an AI-driven Streamlit interface.
MPL-2.0 ★ 1 -
Font discovery and typography tools for AI agents. Explore local open and proprietary collections, refine by visual similarity, compare real text in context, and resolve exact assets through a JSON CLI and MCP. Fonts and model inference stay local; usage rights vary by source.
MIT ★ 1 -
m1nd is a local-first context runtime for coding agents: an evidence-bound, continuously verified model of your code, memory, and change.
npm MIT ★ 24 -
Quiet MCP executor for Claude Code and Codex. Offload edits, tests, refactors, and commits to cheaper local or flat-rate models.
MIT ★ 4 -
postgres-mcp-assistantgithub.com/MeghnaTomar17/postgres-mcp-assistantupdated this year
An AI-powered PostgreSQL assistant built using the Model Context Protocol (MCP). Supports database metadata exploration, secure read-only SQL execution, and natural language querying with local LLM integration.
-
Sub-second local business telemetry & flaw audits in ~210 tokens. Official Model Context Protocol (MCP) server & CLI for Claude Desktop, Cursor, and AI agents.
★ 1 -
Agnostic Local AI Agent with 3-Level Long-Term Memory, Multi-Workspace RAG, High-Performance REST API, and native Model Context Protocol (MCP) server for Claude & Cursor.
-
Run local Ollama models as an MCP server so an agent can send work to a model on your own machine instead of an API.
Apache-2.0 ★ 2 -
Safety-first Home Assistant MCP server and web UI for small local AI models, plus walls for the ha-diorama 3D card. Built with the AI agent HESTIA (Claude Code).
MIT ★ 1 -
@velvetmonkey/flywheel-ideasgithub.com/velvetmonkey/flywheel-ideasupdated this year
The local-first falsifiable decision ledger — MCP server that turns an Obsidian vault into a compounding decision system with multi-model AI council dissent and outcome-driven refutation propagation.
npm Apache-2.0 -
Open MCP server suite that gives coding agents, including local and low-cost models, a verification gate, repo map, LSP, code graph, RAG and knowledge graph, research, checkpoints and watchdogs, with a planner/executor model topology.
MIT ★ 1 -
Unofficial read-only MCP server that lets a local model watch Wyze cameras, remember what they saw and alert you by push, email or text. Vision stays on your machine.
MIT ★ 1 -
MCP control plane for local cognitive labor — job-shaped tools with tiered Ollama models (instant/workhorse/deep/embed), server-enforced guardrails, and measured economics so Claude can delegate bulk work without losing control.
npm MIT