Connect your wearable and health data sources to YOUR ChatGPT, Claude, OpenClaw, Hermes, or any AI that speaks MCP. Sleep, recovery, HRV, workouts, queryable in natural language.
This is the 3rd launch from freddy.. View more
freddy.health
Launching today
Connect your Garmin, Polar, Oura, Withings, Intervals.icu, Hevy, WHOOP, Suunto, Concept2 and more to ChatGPT, Claude, Claude Code, or any AI that speaks MCP. Sleep, recovery, HRV, workouts, queryable in natural language.
Free Options
Launch Team
freddy.
Let's go team. What should we integrate or bring in next?
ApplySeed
This literally can push me to wear my whoop again!!
freddy.
@bengeekly WHOOP is a waitlist right now, but send a chat message from within the app and we will see what can be done about that ;)
Refocus
Querying sleep and HRV in plain language through an MCP client is a neat angle, since most wearable apps still make you read dashboards. I'm building a voice AI that calls aging parents every day, and the thing we keep hitting is how to turn noisy signals into something a family member would actually act on. How do you decide when freddy. should surface a trend unprompted versus waiting for the user to ask?
freddy.
@igorgurovich We leave that choice with the user. What counts as signal versus noise depends on their context, goals, and what they would actually act on. That is the core value behind freddy. Giving people access to their own data so they can explore it with AI, without forcing them into a predefined mold built for someone else. The same applies to unprompted trends, the user should decide what deserves their attention and when. There is no silver bullet that would work for our 15k users.
Dial
the MCP part is what gives me pause here, more than the wearable integration itself. sleep, HRV and recovery numbers going through a general-purpose chat history is a different privacy surface than the wearable's own closed app, that data can end up in logs, in a provider's training pipeline (depending on settings), or just sitting in a chat transcript someone else on a shared account could open. is there a local-only mode where the MCP server never round-trips through a hosted layer, or does every query pass through freddy's own backend before the AI sees it?
freddy.
@galdayan The whole idea behind freddy. is that you can query your data with any compatible AI, from anywhere. It runs on our hosted backend; there is no local-only server mode. Each MCP query goes through freddy. before the requested data reaches your AI client.
Our backend and the AI provider’s chat history are separate privacy surfaces. You are right that data returned to an AI client is subject to that provider’s retention, training settings and account access. Hosting freddy. locally would not, by itself, change that. You CAN use a local AI agent, eliminating that surface.
Our servers handle the source integrations and their credentials. A fully local version would need a different integration model, such as users supplying their own developer credentials where the source allows it. We cannot distribute our application credentials for unrestricted use.