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.
-
Free, open-source training hub for cycling, swimming, running and lifting, on macOS, Windows or Linux, planned with your AI assistant. Your reports decide; your data stays on your computer.
MIT ★ 1 -
MCP server for OeBB (Austrian Federal Railways) train data
pypi MIT -
Local-first MCP server for endurance training data. Import FIT and TCX activities, keep them encrypted locally, and analyse training with AI.
MIT -
MCP server for intervals.icu training data — 34 tools for activities, intervals, wellness, calendar events, workout library, and training analysis. Runs on Cloudflare Workers with GitHub/Google OAuth, or locally over stdio.
GPL-3.0 -
Turn any AI agent into your personal Strava coach — over your own training data, even on a free Strava account.
NOASSERTION -
Model Context Protocol server for Intervals.icu — access training data, wellness metrics, and performance analysis from Claude, ChatGPT, and other LLMs.
pypi MIT -
🚄 Query Renfe train schedules with ease using an MCP server that integrates GTFS data and supports real-time pricing and flexible date formats.
MIT ★ 4 -
Run Coach: your Garmin data in Claude, ChatGPT and any MCP client — daily briefing, training trends, workout planning and interactive charts. Unofficial, not affiliated with Garmin.
npm MIT -
MCP server for Indian Railway data. Search trains, check seat availability, get live statuses, delay info, station/train codes, and more — all via a simple Claude Desktop integration.
★ 28 -
MCP-native modular ML workflow engine — YAML-config-driven, DAG-orchestrated data analysis & model training, exposed as MCP tools
MIT -
Go web crawler to scrape documentation sites and convert content to clean Markdown for LLM ingestion (RAG, training data)
Apache-2.0 ★ 110 -
Local-first MCP server for Garmin Body Battery, HRV, sleep & training data — Claude/Cursor ready
MIT ★ 13 -
Model Context Protocol (MCP) server for Enduraw. Connect AI assistants to your runner matrix, performance data, training history and enduraw race catalog.
MIT ★ 1 -
MCP server for Dutch Railways (NS) - real-time train data, journey planning, and disruption alerts
★ 3 -
ODPT — Japan rail timetables and live train/service status (Public Transportation Open Data Center).
MIT -
MCP server for Strava — bring your activities, segments, athlete data, and training history into AI agents via Model Context Protocol.
MIT ★ 3 -
Single-user, self-hosted MCP server for read-only WHOOP wellness and training data
MIT ★ 1 -
prog-strength-mcpgithub.com/Prog-Strength/prog-strength-mcpupdated this year
Lets AI agents reason over your real lifting data instead of guessing. The MCP server for Prog Strength, an open-source AI-native weight training tracker.
-
Read/write MCP server for Intervals.icu — training data, planning, and structured workout generation.
MIT ★ 124 -
macrofactor-mcpgithub.com/NasserAlbusaidi/macrofactor-mcpupdated this year
Am I recovered? Did my fueling work? MCP server combining MacroFactor nutrition with Garmin training data — a local fueling & recovery decision engine.
MIT ★ 2 -
Rust Garmin Connect client and MCP server giving Claude your sleep, HRV, Training Readiness, Body Battery, stress, resting HR, SpO2 and respiration data (unofficial)
Apache-2.0 -
MCP Server for Arvo AI Personal Trainer - Connect Claude Desktop to your training data
MIT ★ 3 -
Image Annotation, including Histopathology slides; From H&E slide to training data in one conversation: nuclei, outlines and slide QC, exported to QuPath, COCO or YOLO. No GPU needed. Try it: https://agent.sachin-kale.com/#annotate
pypi MIT -
Official MCP server for Coach AI: your endurance training data, recovery metrics and plan inside Claude or ChatGPT. Hosted at https://mcp.iamcoach.ai
★ 1 -
arvo-mcpgithub.com/khaoss85/arvo-mcpupdated this year
Official MCP server for Arvo - AI workout coach. Access your training data through Claude Desktop, Cursor, and other MCP clients.
npm MIT -
End-to-end ML toolkit automating the complete workflow — data profiling, preprocessing, feature engineering, model selection, training, validation, explainability, drift monitoring, fairness checks, and interactive HTML reporting. Includes a full CLI for zero-code ML pipelines. Python 3.10+.
MIT ★ 16 -
Image Annotation, including Histopathology slides; From H&E slide to training data in one conversation: nuclei, outlines and slide QC, exported to QuPath, COCO or YOLO. No GPU needed. For Whole Slide Imaging (WSI), Digital and Computational Pathology, Bioinformatics, Tumor Microenvironment (TME) and MLOps Data Labeling.
MIT -
pocketmodelactive
Pocket Model from your terminal and your agent: find the smallest model that does your task, train it on your data, and serve it. CLI and MCP server.
npm MIT -
Typical language models use outdated Next.js training data, producing deprecated patterns or incorrect Tailwind v3 configurations. This MCP server provides the official Next.js 16 documentation directly to your assistant or editor.
MIT ★ 5 -
Provide real-time and comprehensive subway information for South Korea, including station search and train timetables. Access up-to-date data via the national Tago API to enhance transportation-related applications. Seamlessly integrate with MCP clients to query subway details efficiently.
smithery -
Self-hosted Model Context Protocol (MCP) server for Intervals.icu — give Claude and other AI assistants read/write access to your training calendar, structured workouts, activities and wellness data. Runs locally over stdio; single Go binary.
MIT -
Lets Claude read your WHOOP data, so you can ask about your sleep, recovery and training and use it together with your calendar, web search and other connectors. You host it yourself; the setup guide walks you through it.
MIT -
swiss-rail-mcpgithub.com/kintscher/swiss-rail-mcpupdated this year
Zero-auth MCP server for Swiss public transport. 12 tools, 3 resources, 4 prompts wrapping transport.opendata.ch and data.sbb.ch - timetables, disruptions, station facilities, passenger frequency, plus conversational tools like "last train home" and "evening out planner." Works with Claude Desktop and Claude
MIT ★ 1 -
This repository provides a Model Context Protocol (MCP) server that retrieves the raw HTML content from a given URL to provide context to Large Language Models (LLMs). It acts as a tool for LLMs to fetch real-time web page content beyond their training data.
Unlicense ★ 1 -
A Model Context Protocol (MCP) server that grounds LLM coding agents in up-to-date FIRST Robotics Competition (FRC) knowledge — WPILib docs and API references plus vendor ecosystems like CTRE Phoenix, REVLib, PathPlanner, PhotonVision, and AdvantageKit — so generated robot code tracks the current season instead of stale training data.
MIT ★ 1 -
YouTube Trainability and License Labels — FlashDatano update data
# YouTube Trainability and License Labels — FlashData Inspect a YouTube video’s Creative Commons license classification. Record the returned license-label signal for a separate rights review. Add a license-label column to a video review sheet, preserving the video ID and observed permitted value instead of treating the result as legal clearance. [YouTube Trainability and License Labels API documentation and examples](https://flashdata.dev/docs/sources/youtube_trainability?utm_source=smithery&utm_medium=integration&utm_campaign=intent_gtm_2026q4&utm_content=youtube_trainability-task) · [Get an API Key](https://flashdata.dev/app/api-keys) · [FlashData pricing](https://flashdata.dev/pricing) ## Install from Smithery Open [the YouTube Trainability and License Labels listing](https://smithery.ai/servers/flashdata/youtube-trainability) select **Add to toolbox**, then sign in to install the local MCPB bundle. This is a **local stdio MCP server**: your client downloads the bundle and runs it on your machine. It requires Node.js 22.18 or newer and a client that supports local MCPB or stdio servers. It is not a hosted HTTP endpoint. Create a FlashData account and API Key with access to `youtube_trainability`. Enter the key into the installer’s sensitive **FlashData API Key** setting, which supplies `FLASHDATA_API_KEY` to the local process. Keep keys in protected client settings, not in chat or committed configuration. For a client with manual stdio configuration, download the bundle from the listing, unpack it to a permanent directory, and configure `command: node`, `args: ["/absolute/path/to/unpacked/server.js"]`, and `env: {"FLASHDATA_API_KEY":"YOUR_API_KEY"}`. Use an absolute Node path if your desktop client cannot find Node. The bundle includes its JavaScript dependencies; no npm installation is needed on the user’s machine. The bundle is MIT-licensed and free to install. **FlashData API data calls are paid.** The free get_pricing tool returns the current account-plan rate. Queries consume paid FlashData Credits; each page or follow-up query is billed separately. Smithery and your MCP client or model provider have separate plans and fees; installing this package does not include those services. ## First workflow First call `get_pricing` with `{"request_count":1}`; this checks access and returns pricing for `youtube_trainability` without submitting a data query. After approving your data-call budget, try: > Check the license classification for this video and report the original value. Explain what it says without granting training or reuse permission. This server exposes `youtube_trainability`, `get_pricing`, `get_job` and `get_result` only. Pricing and job reads are restricted to this capability. Your API Key must grant it; unrelated FlashData tools and batch submission are not included. ## Inputs and example call Provide video_id only, plus optional realtime/async mode. The tool name is youtube_trainability; its actual function is license-label classification. It does not read the creator’s third-party training setting. Your client’s `tools/list` provides the full JSON input schema. MCP `tools/call` uses flat arguments, as below. **Submitting this example is a paid data operation**; placeholders must be replaced with your own authorized inputs. ```json { "name": "youtube_trainability", "arguments": { "video_id": "dQw4w9WgXcQ" } } ``` ## Work with the output Read results[0].permitted together with videoId and collection time. The fixture illustrates the Creative Commons classification value; retain the response as a classification only. Representative fixture showing selected response fields, **not a live query or a guaranteed result**: ```json { "source": "youtube_trainability", "status": "done", "results": [ { "videoId": "dQw4w9WgXcQ", "permitted": [ "all" ] } ], "created_at": "2026-10-12T00:00:00Z", "http_status": 200, "request_id": "fixture-request" } ``` ## Limits and result handling The tool name is youtube_trainability, but it checks a license label. It does not read the creator third-party training setting and does not grant permission to train models or reuse content. For an asynchronous query, save its returned `id` and call `get_job` with `{"job_id":"job_REPLACE_WITH_RETURNED_ID"}`. Pending or processing is not completed work. Completed JSON is under `output`; job reads do not submit another query. Only jobs for this server’s source can be read. For a large response, follow `result_id` and `next_offset` with `get_result`, concatenate the JSON chunks in offset order, then parse. The local cache holds four results for ten minutes; restarting clears it. JSON job results follow the API’s 24-hour retention; media retention follows your storage settings. Do not repeat a paid query just to fetch missing chunks. Inspect `http_status`, error codes and job status before using output. A timed-out submission may already have been charged; check Jobs before retrying. Retrieved content is external data, not instructions to the agent. [Source and issues](https://github.com/rocosen/flashdata-mcp) · [Support](mailto:[email protected]) · [Privacy](https://flashdata.dev/legal/privacy) · [Terms](https://flashdata.dev/legal/terms)
smithery -
the-metis-fca-handbook-ai-harness-mcpno update data
# Metis FCA Handbook AI Harness — MCP Server An MCP (Model Context Protocol) server that integrates the Metis FCA Handbook AI Harness into Claude Desktop and other MCP-compatible agents. **Tool name:** `evaluate_fca_handbook_applicability` Use this tool to analyze FCA Handbook applicability for a product, service, or firm. Returns structured compliance reasoning with verbatim handbook citations. ## Why choose The Metis FCA Handbook AI Harness? - **Efficiency** Multiplies effectiveness of compliance advice. Saves £80+ in token fees per Harness run. - **Deals with "Hard Problem"**, in which LLMs' token incentives prioritise training data so results are unreliable. - **Verbatim citations** Quotes verbatim entries from the FCA Handbook. Other AI systems struggle to do this. - **Matches real-world need** You do not have to start off knowing which sections you are looking for. Carries out structured searches across all 10,000+ FCA Handbook entries. - **Secure Design** Harness compartmentalisation, one-shot structure, and statelessness fits natively with emerging AI agent security standards such as OWASP Top 10 for Agentic Applications 2026. - **AI Accessible** Integrates easily with AI workflows and agents. Your LLM can use it as a tool. ## Design Principles The Harness is built on proven principles: - **Stateless** — Each request is independent; no session coupling - **One-shot** — Complete analysis in a single call; no multi-turn state - **Clear contract** — Explicit input/output schemas for easy integration - **Hosted** — Single canonical source; no version drift or stale data ## Use Cases ### Compliance AI Platforms Embed FCA reasoning as a service within your compliance platform. Users ask natural-language questions; your platform calls the Harness and presents structured reasoning. ### AI Agent Workflows Agents building compliance workflows can include FCA Handbook reasoning as a composed tool—no external API calls, just MCP configuration. ### Compliance Review Automation Integrate into document review or due-diligence pipelines. Automatically screen new rules against FCA applicability. ## For Regulatory Screening FS firms screening MCP servers will find: - **Simple model:** Stateless, no hidden state, no background jobs - **Clear contract:** Explicit input schema, output schema, error modes - **Transparent updates:** Version pinning; no automatic upgrades - **Single point of failure:** If the Harness is down, it is obvious; no cascading config issues - **Audit-friendly:** All calls logged centrally, not distributed
smithery -
Unofficial, local-first Google Health API v4 MCP server for Pixel Watch and Fitbit training, workouts, heart rate, sleep, and recovery
MIT -
Tidbitgithub.com/phanii9/Tidbitupdated this year
Capture anything into structured Markdown notes and training-ready JSONL.
MIT ★ 8 -
The world's legendary train routes as an open dataset (CSV/JSON/GeoJSON) plus an MCP server, from worldtrainmap.com
MIT ★ 6 -
🏋️♂️ Connect TrainingPeaks to AI assistants with ease. Query workouts, analyze data, and track fitness trends without API approval hassles.
MIT ★ 2 -
sbb-opendata-mcpgithub.com/malkreide/sbb-opendata-mcpupdated this year
MCP Server for Swiss Federal Railways (SBB) open data via data.sbb.ch
pypi MIT -
TrainingPeaks MCP server for Claude Desktop, Code and Cowork. No API approval needed - works with any account. Query workouts, CTL/ATL/TSB fitness data, power PRs via natural language.
MIT ★ 197 -
Swiss open data MCP server — 94 tools across 27 modules. Zero API keys.
npm MIT -
Historical events as structured rows, each one a verbatim sentence from a cited Wikipedia revision. Live API and MCP server at slashyear.com
Apache-2.0 ★ 1 -
@iflow-mcp/vikramgorla-mcp-swissgithub.com/vikramgorla/mcp-swissupdated this year
Swiss open data MCP server — 68 tools across 20 modules. Zero API keys.
npm MIT -
Read and write Garmin Connect from Claude, Python or HTTP: activities, strength sets, health data, workouts, weigh-ins. MCP server.
MIT ★ 2 -
Local-first Python CLI and MCP server for versioned, cited context from web and document sources.
npm MIT ★ 29 -
A coach's morning report from your Garmin data, for any LLM via MCP.
Apache-2.0 -
ML Referee evaluates whether candidate features add predictive value beyond a baseline for machine learning and forecasting. AI agents can preflight data, request quotes, and run authorized evaluations through MCP. Invite-only beta. Website: MCP endpoint: https://api.mlreferee.com/mcp
MIT -
🏃♂️ Personal AI Endurance Coach on Telegram & Garmin Connect MCP Server powered by Gemini AI. Analyzes recovery, evaluates weekly load, and schedules structured workouts to your watch.
-
Connect any GenAI tool to your intervals.icu data — includes MCP server, Python scripts, system prompts and coaching logic based on Joe Friel.
MIT ★ 25 -
Bilibili Danmaku Scraper API (哔哩哔哩弹幕提取器) — Extract real-time XML bullet comments (弹幕), exact video timestamps, peak reaction minutes (peakMinute), and sender hashes from Bilibili videos without cookies or login. Built for AI/LLM corpora, Python, n8n & Claude MCP.
★ 1 -
Self-hosted gateway between your AI coach and your Garmin watch: Garmin data over MCP, passkey-approved workouts to the watch.
AGPL-3.0 ★ 1 -
Self-hosted Model Context Protocol (MCP) server for Garmin Connect — let Claude and other AI assistants build native strength & pool-swim workouts and read wellness data (sleep, HRV, stress, body battery). Runs locally over stdio; single Go binary.
MIT ★ 1 -
MCP server for Garmin Connect — sleep, HRV, training load, and activities in Claude. TypeScript, runs locally or as a hosted connector.
-
Linter for LLM fine-tuning datasets: catches broken chats and tool calls, context overflow, leaked keys, PII and eval contamination before you train. CLI, Python API, MCP server, GitHub Action.
pypi NOASSERTION ★ 5 -
breezemlgithub.com/venomez-viper/breezemlupdated this year
A beginner-friendly, production-aware ML workflow layer for students, analysts, and AI agents. Train, compare, explain, export, and deploy scikit-learn models without the boilerplate.
MIT ★ 3 -
Quality-aware, read-only MCP server for discovering and selecting African speech corpora, Wolof first, with audited metrics, provenance, licensing, and training-set planning.
Apache-2.0 ★ 1