Autotune is a great system for analysing the data stored in Nightscout to recommand changes to basal, insulin:carb ratio and insulin sensitivity factor settings to improve blood glucose control, but as I don't use an OpenAPS rig I found it a hassle to have to manually create a compatible profile.json file each time I wanted to run it manually. All my profile data was already in Nightscout, so I created AutotuneWeb to convert the Nightscout profile to the OpenAPS format.

I could also never remember the exact command line to run Autotune, so I also got AutotuneWeb to run it for me.

In the pay-it-forward nature of #WeAreNotWaiting, I then opened up AutotuneWeb to help anyone else with Nightscout get the same results.


AutotuneWeb is using basals, ISF, or carb ratio that are not my current settings.
AutotuneWeb pulls from your Nightscout Profile. Make sure your Nightscout Profile is up to date with the settings you would like AutotuneWeb to compare from when it runs. If your current Nightscout profile does not accurately reflect the settings you're using, the suggestions generated by Autotune may be misleading.
I get a 401 error below my URL and it doesn't run.
AutotuneWeb pulls from your Nightscout. Nightscout must be set with "readable" in AUTH_DEFAULT_ROLES in order for your Nightscout to be readable by AutotuneWeb
I get a warning about missing "rate" property in my temp basal data
This is an extra bit of information that Autotune expects in the Nightscout data that isn't always added. If this data is missing then Autotune won't have a full picture of how much insulin has been delivered and therefore the recommendations will be out and should not be used.
I haven't received my results by email yet
It can take 10 - 20 minutes to run, so be patient! Check your spam folder to make sure the results haven't been blocked.
The recommendations don't look right
I'm not an expert on the interal workings of the Autotune program itself, and you should get the same results running Autotune here as if you did it youself. If the results don't look right, please don't use them and reach out on the Facebook groups for advice
Where can I learn more about Autotune?
The official documentation is on the OpenAPS site.
What do the colours mean on the results email?
Any major changes suggested by Autotune are highlighted with either a yellow background (for changes above 10%) or a red background (+20% or -30%). While these suggestions may be correct, the highlights are there to draw your attention to them. Please exercise appropriate caution when making any changes to your settings based on these results.
What about the green & red blocks next to the basal suggestions?
In order to produce basal suggestions, Autotune needs to see BG data for each hour where any fluctuations are not due to other factors such as carbs. If all the BG data for an hour is dominated by other factors, it instead produces a suggestion based on averaging the basal rates for the surrounding hours. The green blocks indicate how many days Autotune used fluctuations in the BG data to produce the suggestion, and the red blocks indicate how many days it interpolated a result based on the surrounding hours.
Why is the distinction between basal suggestions based on BG fluctuations and interpolating other basal rates important?
If you have a basal profile in which you have a different basal rate at a time of day when you normally have carbs on board as well, Autotune will tend to flatten out that different basal rate and compensate with a corresponding change to the carb ratio. Depending on the underlying reason for your different basal rate, this may be correct but also may cause it to suggest a lower basal rate that will lead to high BGs at the affected time and also a stronger carb ratio that may lead to lows at other times. If you see a lot of red blocks in your basal suggestions, please take extra care before enacting any changes based on the information in the results.
How can I see what this site is doing with my data?
All the code for this site is open source, so feel free to look at the GitHub repo to see exactly what it's doing. You can also take a look at the privacy policy
What version of Autotune is it running?
The details of the version of Autotune used on any particular run is included in the result email. The latest job ran with version 5313363b
Why do I get the same results regardless of whether I tick "Categorize UAM as basal"?
If Autotune detects at least 1 hour of carb absoption, it will assume that all carbs have been entered and so you get the same result as if this option had been ticked even if it wasn't.