OpenAPS – Still changing the diabetes world…

OpenAPS – Still changing the diabetes world…
OpenAPS – Still changing the diabetes world…

If you don’t remember, back in May 2017 (some two and a half years ago), Dana wrote a post on her blog with the title “Choose One: What would you give up if you could?”. She asked:

How many of these 5 things do you think are possible to achieve together?

  • No need to bolus
  • No need to count carbs
  • Medium/high carb meals
  • 80%+ time in range
  • no hypoglycemia

And went on to note that with OpenAPS oref1, four of the above five are possible, and in some circumstances, five out of five. Which I’ve reminded myself about this week having decided to investigate some of the theories relating to Fiasp that I’ve heard over the past year. 

But the thing that stands out from this is that, while we wait for commercial offerings to even make it to market and let’s be realistic here. There are still only two approved systems available at time of writing in November 2019, Diabeloop and Medtronic’s 670G, with an expected soon on the Tandem Control:IQ enabled T:Slim imminent, none of them provide that “4/5” performance. 

Admittedly, they include other features that none of the DIY APS systems do, namely, simple set-up and auto-configuration, but none currently quite match the day to day performance that DIY systems seem to (regardless of whether the ability to deliver a single correction bolus once per hour under specific circumstances stratifies an algorithm as more advanced than only using basal).

In that sense, OpenAPS’s core algorithm is probably the most advanced, easily accessible automated insulin dosing algorithm that the general public can get their hands on. And that’s more than two years after it was released. 

What’s your point?

Why am I posting this now? Well version 0.7.0 has just been released, providing some welcome operational and stability updates, as well as a number of functional new features:

  • The ability to manually autotune for weekdays and weekends separately;
  • The dev functions for adjusting insulin peak time and insulin end time in autotune have made it into production
  • Autotune on the rig reflects the web version, showing how often the value is interpolated versus tuned 
  • Support for a custom SMB interval
  • The ability to set a custom bolus increment
  • x12 users no longer need manual set-ups
  • Support for additional treatment entry methods
  • Support for retrospective autosens
  • Allow high_temptarget_raises_sensitivity for temp targets >=101mg/dl and exercise mode > 100mg/dl (vs 111 and 105 previously).
  • Ignore glucose values that are older than the last calibration record to avoid dosing for jumps following a calibration
  • If CGM data is changing less the 1mg/dl/5m for 45 mins, then cancel high temps and shorten any long zero temps
  • autosens-history.js
  • Sample json file for use with oref0-simple-simulator
  • Updates to bluetooth and online scripts to improve performance
  • Significant updates to the core pump comms processes to use Go
  • A lot of syntax fixing and code refactoring to improve efficiency
  • Addition of xdrip-js to the set-up script to allow the rig to act like a dexcom receiver and operate standalone.

For details of all of these, go to

A lot of updates, undertaken by a lot of people, all of whom I’d like to publicly thank for their work, as it makes my life easier…!

One more thing…

And in the meantime, as I recollect on the fact that I’m now well into my fourth year of using a DIY closed loop, and the associated benefits I’ve seen with it, it’s worth considering that Dana’s loop has been closed for nearly five years (five years in December I believe).

Five years of forcing commercial systems to step up to the plate. Five years of causing regulators to re-asses how they deal with Diabetes Technology. Five years of leading the charge to make the lives of thousands of people better. 

Five years of changing the diabetes world.

Dana, let me raise a glass to you!


  1. Well done OpenAPS! Thanks to all the pioneers! I chose Loop almost a year ago, and have never done so well.

  2. I think we need another type of insulin if we wish not to bolus and count carbs. Unless we can inject the insulin into the portal vein. Present subcutaneous insulis just are not fast enough.

  3. Here here. Not on the OpenAps train yet – still MDI – but this work is outstanding. Thanks for sharing.

Leave a Reply

Your email address will not be published.