Agree, HRV/RHR comments have been already turned off for those not pulling the data (myself included 
). At the next push you will notice the changes.
Thank you all and keep the feedback coming. In the next 10 days or so the final comment for release will take shape.
Intervals.icu imports data from HRV4Training, without an API - I have HRV4Training set to save to a folder in Dropbox, and Intervals uploads it from there. After filling out the daily questionnaire in HRV4Training, it’s only a couple more taps to do this. The import is just parsing a csv file.
As I recall, the iOS version of HRV4Training saves the file slightly differently than the Android version that I use.
Noted thanks @Robert and added to our ticket on this for consideration.
Hi @Prof … can you add me to beta on my athlete account to test out features as well. Thanks!
For sure @mgiven, however the email associated with this account goes to a coach account so I’ll need more info please… feel free to DM me.
Yes, I export / import to Intervals now and again.
The about file says this
This export is of all of your raw RR interval data.
Every reading that you have ever taken is included in this folder. Each file is an individual reading and is named by date and time of that reading.
Within the file is a line separated list of RR intervals. These can be uploaded into a software like Kubios for further analysis.
I’ve had this and it’s not so much that sessions were unplanned. I might often record a main session separate from the warn up or cool down. Which means there maybe one planned session with two or three actual sessions. The total load is about right and the intervals I did as best.
It would be better in my view if the AI avatar took a more holistic view at the overall load executed each day and over the week. To give a view of actual load vs. planned and progressive overload.
So I don’t know if it’s the coach avatar / beta glitch, or just the existing fully released AI feedback, but said coach is acting like a drunken sailor.
Your current plan concludes tomorrow, so think about selecting one of Athletica’s personalized plans to continue your progress.
No it doesn’t. In fact it’s only just entering week 2.
You executed your planned aerobic run with admirable compliance, completing 66 minutes instead of the planned 50, which shows your dedication!
Like others have experienced, 66 is actually the load, not the duration.
I’m not an expert on AI, but based on some general stuf I’ve read and my own experience using large language models to try to summarize financial stuff, they are generally not great with numbers. They get confused very easily. I wonder if the AI description is just an application of one of those models to the table spit out by athletica based on the incoming Garmin file.
I have the same about plan ending, plus the RHR comments are back. Thought the latter had been turned off?
Edited to say, the Avatar also doesn’t have anything to tell me about what is scheduled next week.
I think the AI avatar is getting a bit ahead of itself
Get ready for the ultimate test of your cycling skills, where you’ll tackle everything from mountains to mud puddles all in one day
Wondering what the mud puddle training sessions will be !
Bring on the mud puddles!
If the coach is utilising one of the LLMs (which I assume it may be), just wait until he advises violent solutions or refuses to say certain completely random and innocent words 
The AI avatar got confused between watts and load today. The numbers are correct, just the added units incorrect.
Hi! Little feedback from me 
I really like the AI summary feature, but I think the usability could be improved with a different approach. Here are my thoughts:
Don’t get me wrong—I love the summaries and that the information is available! I just think these changes could make the experience even better.
It would be good if the coach analysed executed cadence and power for strength endurance workouts.
How well did I comply with the cadence targets, and how was my power for the intervals against target?
