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.
-
tuningengines-cligithub.com/cerebrixos-org/tuning-engines-cliupdated this year
Tuning Engines CLI, MCP server, and Python agent runtime adapters for governed model, agent, skill, and MCP workflows. Fine-tune open-source LLMs, run inference, manage datasets/evaluations, and connect LangGraph or Temporal while Tuning Engines handles p
npm MIT -
learn-mcpgithub.com/fvegiard/learn-mcpupdated this year
AI Engineering Learning Reference MCP Server — 550+ searchable references for MCP, RAG, agents, fine-tuning, MLOps, prompt engineering, inference, vector DBs, GPU compute, autonomous agents, and Claude Code workflows
MIT -
MCP context engine for AI coding agents: a local code graph answers "which files does this task need?" — folded evidence packets instead of whole files. Recall 1.000 on repos it was never tuned on, 45% fewer context tokens in a real Claude Code session. Rust, local-first, MIT.
MIT ★ 5 -
Open-source text-to-SQL engine you tune and measure on your own database
Apache-2.0 ★ 81 -
ml-mcpgithub.com/PushPullCommitPush/ml-mcpupdated this year
A comprehensive MCP (Model Context Protocol) server for ML model training, fine-tuning, and experimentation. Transform your AI assistant into a full ML engineering environment.
NOASSERTION ★ 4 -
C++20 trading-engine framework for Gate.io — 9 layers, 1054 tests, JSON-RPC/MCP control plane, Qt client. AI tuning is experimental and unverified.
GPL-3.0 ★ 5 -
AI motion design studio for macOS. Describe a video, the agent builds it, you polish it on a timeline. Render engine tuned for Apple Silicon.
-
Local-first, no-API-key MCP server for web search, fetch, and document reading. Multi-engine (DuckDuckGo, Mojeek, Startpage), smart Playwright fallback, FTS5 cache, LLM-tuned output.
pypi MIT ★ 183 -
Make your model serve faster on your GPU, and prove it. Skills for Claude Code, Codex and OpenCode, plus a CLI and MCP server: your agent picks and tunes inference engines, can write GPU kernels, and backs every number with client-timed benchmarks, anti-cheat gates and quality checks.
pypi Apache-2.0 -
Graph-Augmented Code Intelligence RAG engine pairing deterministic AST call graphs with dense vector search to boost direct-callee Recall@8 from 0.0% to 53.3% (+53.3 pp) under a 5.8ms p95 latency bound. Powered by a QLoRA fine-tuned Qwen2.5 adapter (70.02 CodeBLEU), quality verification harness, and production Model Context Protocol (MCP) server.
MIT -
Vedakshano update data
Astronomical ephemeris and Vedic astrology (Jyotish) engine, written in Rust and exposed as 17 MCP tools: natal charts, panchanga, Vimshottari dashas, vargas, transits and transit search, muhurta search, ashtakavarga, shadbala, karakas, combustion, drishti, bhavas, synastry and composite charts. Runs locally over stdio. No network calls, no telemetry, no API keys and no configuration. **Install:** `cargo install vedaksha-mcp` then `vedaksha-mcp --stdio`, or `docker run --rm -i ghcr.io/arthiqlabs/vedaksha-mcp:v7.2.0 --stdio` (amd64 + arm64). Positions come from a clean-room implementation of VSOP87 and ELP/MPP02, with optional JPL SPK kernels, validated against JPL Horizons on a committed 24,350-row fixture; the SPK path agrees to 0.103 arcsec. Eleven sidereal ayanamsha systems are each derived forward from a primary source rather than tuned to match another implementation, and the derivation for each is recorded in the repository. **Licence:** BUSL-1.1 (source-available). Free for any non-commercial use. Commercial use is a one-time USD $500 per organisation. Each version converts to Apache-2.0 four years after its release.
smithery -
fine-tuning-osgithub.com/Casius999/fine-tuning-osupdated this year
Zero-Data fine-tuning operations MCP server - 64 tools for the full LLM fine-tuning delivery lifecycle, callable by Claude Code
Apache-2.0 -
MCP-native multi-channel outreach server. Sequence orchestration, reply classification, and auto-tuning as typed MCP tools.
MIT -
rag-eval-enginegithub.com/aptsalt/rag-eval-engineupdated this year
Production RAG with continuous evaluation — hybrid retrieval (BM25 + vector/RRF), LLM-as-judge scoring on faithfulness/relevance/hallucination, MCP server, adaptive retrieval tuning. 68 tests.
MIT -
agentforge-mcpgithub.com/sciscale-studio/agentforge-mcpupdated this year
MCP server for AgentForge — turn rough requests into tool-tuned, quality-checked prompts for AI coding agents.
npm MIT ★ 2 -
copilot-studio-simulatorgithub.com/butelo/copilot-studio-simulatorupdated this year
Unofficial local Copilot Studio simulator: declarative agents (topics, tools, MCP, code execution, knowledge, child agents) + evals + auto-tuning, built on the pi harness SDK
MIT ★ 1 -
An agent-native MCP for composing, tuning, rendering, mixing, and auditing native VOCALOID3/4 projects — built just for fun.
MIT ★ 14 -
Wake Intelligence: 5-Layer Temporal Intelligence for AI Agents A production-ready Model Context Protocol (MCP) server implementing a temporal intelligence "brain" with five layers: Past (causality tracking), Present (memory management), Future (predictive pre-fetching), Adaptive (meta-learning — per-project weight tuning), and Personality (tempora
npm MIT ★ 4 -
MCP server for VRoid Studio. Lets AI agents create and edit characters in the editor — choose presets, tune sliders and colors, frame the camera, and save. Drives the editor's own API through a BepInEx/IL2CPP plugin: no mouse or OCR automation; works with the window minimized
MIT ★ 14