I’ve been using the Aqara FP2 for sleep monitoring for years. How can I export my sleep data/reports into ChatGPT?
You could use the Aqara MCP Server, which can access the logs directly from the cloud. However, it only goes back a maximum of 7 days, as the data isn’t kept any longer than that. At least that’s what I’ve noticed with my sensors.
Since standard cloud-based AI’s usually don’t have access to this, you would need to install a local AI application like Cursor (you can download it here: Cursor · Download) on your PC or Mac. You can find the credentials and instructions on how to set up the MCP server in Cursor here:
Otherwise, as far as I know, there is currently no built-in method to export local logs directly from the M3 hub, or am I mistaken?
There is also the theoretical option of writing a custom script using the official Aqara Developer API. Unfortunately, from what I’ve seen recently, they aren’t approving new developer accounts at the moment, so that’s probably not a viable option. And even if it worked, the API logs are likely still restricted to that same 7-day window.
If you happen to run Home Assistant and your sensor data is logged there, that would be by far the best option for long-term storage and export. But I’m guessing that’s not the case for your setup, right?
1 Like
Thanks for following up on the FP2 sleep tracking. I noticed an inconsistency between the sleep logs and the detailed reports. The logs only list basic status changes (asleep, awake, left bed), but the main reports show deep sleep and awake times. Since overall sleep duration and deep sleep metrics are key daily health indicators.I will try to get sleep report instead of sleep logs.
OK, I understand. Unfortunately, I only use the zone detection feature on the FP2, not the sleep tracking feature. I’m therefore not sure if you can access the reports you mentioned using AI with either the MCP Server or the Aqara Agent Skills. You can try it and see what happens.
@markjyu Out of curiosity, does the sleep tracking feature work well even when you’re under a blanket?