Code is hosted here: github.com/SweatStack/s...
Currently the server needs to be run locally (instructions in the README). A hosted version might follow later.

The SweatStack MCP server built using the official Python SDK for MCP servers: github.com/modelcontext...
...and uses the SweatStack Python library: developer.sweatstack.no/getting-star...

Even with Claude Desktop's code execution limitations, it can already analyze basic workout data from SweatStack. This proof-of-concept shows a path forward: As AI tools evolve, tools like this will become proper data analysis assitants for anyone interested. Small steps!

The SweatStack Python library has built-in support for Streamlit, making it incredibly easy to start building your dream app.
developer.sweatstack.no/learn/integr...

Currently works with Garmin Connect and @intervals.icu. More integrations will follow.

Try it out at sweatstack.run 🚀

Express Login gives you access to your data in Python in under 30 seconds at SweatStack.run!

An example of code/sandbox sharing: sweatstack.run/store/3232be...

Usecase is similar but different from the SweatStack Console: For quick explorations, the Console is slightly nicer to work with because it's a proper REPL that keeps a Python session running. However, for larger scripts that you want to keep and/or share, sweatstack.run is the way to go.

It automatically authenticates with your SweatStack account using OAuth2/OpenID and installs any dependency you import.