M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
C

Coach Watts

Community
Open-source AI endurance coach and unified fitness dashboard
How to install ↓ Compare GitHub source repository ↗
Category
Other #230 of 230
Stars
★ 100 Popular
Transport
Streamable HTTP
Runtime
Hosted (no local runtime)
Credentials
API key / credential required
License
Apache-2.0
Last commit
9FMRS · D

Coach Watts offers a robust open-source AI coach platform focused on endurance sports, with extensive features like AI analysis, nutrition tracking, and multi-platform integration. The MCP server allows remote access to data via HTTP, providing seamless integration for technically capable athletes. However, it requires significant setup and careful data management.

Strongest · Maintenance 3/20 Weakest · Security and permissions 1/20

Reliability
2/20
Security and permissions
1/20
Maintenance
3/20
Documentation
2/20
Setup experience
1/20
Why each score
Reliability 2/20
Only the manifest and README were provided; there is no source code, no CI, and no test evidence. The manifest repository points to hdkiller/coach while the object under review is watt-mind/coach, which is a concrete inconsistency. The README describes the product dashboard, not the MCP server's startup, tools, or behavior. The init/tool-list handshake cannot be verified, so a conservative score of 2 is given.
Security and permissions 1/20
No source code was supplied, so credential storage, least privilege, data boundaries, and external network behavior cannot be audited. The OAuth 2.0 claim in the manifest is unverifiable. The README contains no real secrets or tokens, but it also does not disclose permission scopes, data-flow, or a security-response channel. The repository identity mismatch adds further uncertainty. No definite red-line issue was found, but the lack of evidence is a major deduction, hence 1.
Maintenance 3/20
The project has an Apache-2.0 license and is not archived, but the only signals are stars, 45 open issues, and a self-reported version. There is no commit history, release timeline, dependency update evidence, or issue response pattern. Stars and issue counts are discovery signals only and do not earn points, so maintenance is scored 3.
Documentation 2/20
The README explains the fitness platform, self-hosting, and integrations, but it does not document the MCP server's tools, parameters, authentication, limitations, or troubleshooting. The linked coachwatts.com MCP documentation is outside the supplied source evidence. The documentation does not cover the reviewed server capability, scoring 2.
Setup experience 1/20
Only a remote HTTP URL and self-hosted application setup steps (Node 24, Docker, Google Cloud) were provided. There are no mainstream MCP client configuration examples and no clear steps to connect a client to this server. The install path cannot be reproduced or verified from the given files, so setup is scored 1.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the network

Best for

  • Endurance athletes (cycling, running, triathlon) seeking professional-level insights.
  • Users who prefer self-hosting to have full control over their data.
  • Athletes looking for AI guidance that considers a holistic view beyond just power numbers.

Not for

  • Casual users who do not want to deal with technical setup.
  • Individuals who want a simple step counter without multiple platform data collection.
  • Users seeking advanced strength training plans for gym workouts, rather than endurance focus.

Required permissions

  • Access to user-authorized third-party fitness platforms (e.g., Strava, Whoop, Oura) to sync data.
  • Authentication via OAuth 2.0 to securely access user data.
  • Storage and processing of personal health and training data on a self-hosted server or cloud.

Risks and side effects

  • Privacy risk: Aggregating health data requires access to multiple third-party accounts; if credentials are compromised, there is a risk of data breach.
  • Data accuracy: Analysis depends on the accuracy of synced data; sync issues may lead to incorrect recommendations.
  • AI errors: AI analysis powered by Gemini may have inaccuracies or misinterpretations, so it should not replace professional medical or coaching advice.
  • Dependency on external services: AI and some integrations rely on Google services and third-party API keys; outages may affect functionality.

Setup

Before you start

Runtime：Hosted (no local runtime)

According to the README, the server is served as a remote MCP server at https://coachwatts.com/mcp. To use this MCP server, you would configure the remote URL in a client that supports remote servers. However, no specific client configuration example is provided. The full self-hosted setup requires cloning the repository ([email protected]:hdkiller/coach.git), setting up environment variables, running Docker containers for PostgreSQL, and following the quick start instructions to run migrations.

Check that it works

Add the remote MCP endpoint https://coachwatts.com/mcp (streamable-http) in your client, complete OAuth 2.0 login, and the tool list should show Coach Watts training/nutrition/wellness tools; asking for a recent workout analysis confirms the connection.

Troubleshooting

  1. Ensure Node.js 24, Docker, and Google Cloud account are set up.
  2. Check that .env file is correctly populated with API keys and database credentials.
  3. Run `docker-compose up -d` to start PostgreSQL and ensure port 5439 is accessible.
  4. Run `pnpm prisma:generate` and `npx prisma migrate dev` to set up the database correctly.
  5. If the remote server is not connecting, verify your client supports streamable-http remotes and that the URL https://coachwatts.com/mcp is reachable.

Things to try

Once connected, you can ask your AI assistant things like:

  • Analyze my latest cycling workout and give me an execution score
  • Based on my HRV and sleep data, should I push hard or rest today?
  • Generate an adaptive training plan that fits my schedule

Use cases

Aggregate training, nutrition, and wellness data into a unified dashboard.
Receive personalized training suggestions and daily recommendations.
Analyze past workouts with AI-driven insights.
Generate adaptive training plans based on schedule and goals.
Ask training-related questions directly to the AI coach and receive data-backed answers.

Overview

Coach Watts is a comprehensive, self-hosted endurance coaching platform designed for cyclists, runners, and triathletes. It acts as your 'Digital Twin,' aggregating data from your favorite fitness platforms and using Google Gemini AI to provide professional-level analysis, personalized training plans, and daily recommendations. It understands context—analyzing not just power numbers, but also recovery, sleep, nutrition, and life stress to guide you toward peak performance. This MCP server provides access to training, nutrition, wellness, and performance data via a remote endpoint, secured with OAuth 2.0.

Similar servers

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 659 · Tools 11 Compare with this →
Compare with other MCP servers

Source revision 4d8d65c297cb Data synced 2026-10-19 Read the FMRS scoring method