PEOPLE WITH DIABETES BUILT THEIR OWN ARTIFICIAL PANCREAS BECAUSE THE APPROVED ONE WAS YEARS AWAY
Patients reverse-engineered insulin pumps and built a closed-loop glucose control system rather than wait years for approved versions.
HealthOpen-hardware
- difficulty
- ●●●●○
- time
- a weekend-plus, then ongoing tuning with your clinical team
- license
- AGPL-3.0
- repo
- repo ACTIVE1,142 stars
●●●●○ · a weekend-plus, then ongoing tuning with your clinical team · AGPL-3.0 · 1,142 stars · repo ACTIVE
WHAT YOU’LL NEED
- 3D printer + filament — printable parts — files are in the repo
- Electronic parts — full list with part numbers in the repo BOM
- Dev board / microcontroller — runs the project firmware
Partner
COMPAREE VERDICT
This is a patient-built closed-loop insulin delivery system: a Raspberry Pi Zero with a soldered radio module, a battery, a 3D printed case, and a small purple board called RileyLink that translates between your phone and an insulin pump. The software reads a continuous glucose sensor every few minutes, predicts where blood sugar is heading, and adjusts insulin delivery without anyone pressing anything. The whole thing was reverse-engineered by patients and parents who published the pump radio protocols under the banner #WeAreNotWaiting. The documentation is thorough, the community is active, and the results are real — thousands of people run this. The single biggest barrier is not the build itself; it is that you need a compatible insulin pump, and the models that work (older Medtronic pumps with unencrypted radio) are discontinued and expensive on the second-hand market. You also need a compatible CGM, a clinical team who will work with you on this, and the willingness to tune basal rates, insulin sensitivity factors, and carb ratios yourself. This is not a medical device. It is a do-it-yourself system built by patients who decided not to wait for approval, and every person who runs it does so at their own risk. If you have type 1 diabetes, an endocrinologist who will support you, and the determination to learn how closed-loop control actually works, this is as real as it gets. If you are looking for a consumer product, wait for the approved ones.
IN THE REPO
NOT IN THE REPO
- —AndroidAPS is the Android app that closes the loop. The hardware rig (OpenAPS) is a separate repository with full build instructions, Gerber files for RileyLink boards, and 3D printable cases.
- —This is not an approved medical device. Every person who runs it does so with their own clinical team and accepts full responsibility.
- —The licence is MIT for software (AndroidAPS, Nightscout) and hardware (RileyLink). Commercial use is permitted, though most users are patients building for themselves.
- —You need a compatible insulin pump (older Medtronic models that use the unencrypted radio protocol) and a compatible continuous glucose monitor. The pump compatibility list is in the docs and shrinks every year as models are discontinued.
- —The repository is actively developed as of August 2026. The broader #WeAreNotWaiting movement spans OpenAPS (reference design), AndroidAPS (Android app), Loop (iOS app), and RileyLink (radio hardware).
- —Safety documentation is extensive and assumes you understand diabetes management. This is not a set-and-forget system; it requires tuning, monitoring, and clinical oversight.
Can I build this?
Build at your own risk. Projects involve tools, electronics and sometimes mains voltage — follow the creator’s safety notes.
Partner · KickstarterHeyGears G1: 10M+ colors and transparent parts in one print, plus UV printing on flat objects. Figures, parts and labels — no painting.
Gallery
Start here
Navigation into the creator’s own docs — we don’t rewrite the guide, we route you to the source.
- 1.Read the AndroidAPS documentation front to back before buying anything (The docs assume you already understand diabetes management. If terms like basal rate, ISF, or carb ratio are new to you, this is not the right project yet.)
- 2.Check the pump compatibility list and secure a compatible pump(Most compatible pumps are discontinued Medtronic models. Expect to pay several hundred dollars for a used one, and verify firmware version before buying.)
- 3.Join the community (Facebook groups, Discord, or the AndroidAPS Google group) and ask questions(The learning curve is steep. The community is experienced and helpful, but you must do the work yourself.)
- 4.Build or buy a RileyLink board(You can order a RileyLink from community members or build your own from the Gerber files. Either way, you need this to translate between your phone and the pump.)
- 5.Set up Nightscout for remote monitoring(Nightscout is a separate project that hosts your glucose data in the cloud. It is optional but widely used.)
- 6.Work with your clinical team to set initial parameters, then tune in open-loop mode before closing the loop(Do not enable automated insulin delivery until you have verified every parameter in open-loop mode and discussed the plan with your endocrinologist.)
Resources
Documentation, files and community threads for this build — we link straight to the original sources and never rehost the creator’s files.
Partner · KickstarterCORE POWER BANK S: 0.2" thin, 2.8 oz, 5000 mAh with 22.5W fast charging. Semi-solid battery, carbon fiber, IPX7 waterproof.
KNOWN ISSUES
- The biggest trap is not the build — it is finding a compatible insulin pump. The Medtronic models that work are discontinued, and the second-hand market is expensive and full of firmware versions that will not work. Verify compatibility before you buy.
- This is not a project you build for someone else. It is a medical system that requires the person using it to understand how it works, how to tune it, and how to intervene when it does not. If you are building it for your child, you still need to understand it yourself.
- The documentation assumes you already know diabetes management. If you do not understand basal rates, insulin sensitivity factors, or carb ratios, you are not ready for this. Take a diabetes education course first.
- Closed-loop control is not set-and-forget. You will spend weeks tuning parameters, watching overnight graphs, and adjusting settings. This is normal and required.
- Your clinical team must be willing to work with you on this. If your endocrinologist is opposed, do not proceed — you need someone who can help you troubleshoot when something goes wrong.
- The build itself is straightforward if you can solder and follow Linux instructions. The hard part is the weeks of tuning that follow, and the responsibility of running a system that delivers insulin based on code you control.
Is this legal?
Yes. It is not an approved medical device, but it is legal to build and use for yourself. You cannot sell it as a medical device or present it as approved.
How much does it cost?
The hardware rig is under $100 if you already have a compatible pump and CGM. The pump is the expensive part — expect several hundred dollars for a used compatible Medtronic model. CGM sensors are a recurring cost.
Can I build this if I do not have diabetes?
No. This is a system for people with type 1 diabetes who already manage their condition and want tighter control. Do not build it for someone else unless you are their parent or caregiver and they cannot build it themselves.
What happens if the rig fails?
You fall back to manual pump control, exactly as you did before. The pump still works normally; the rig only sends it commands. Keep backup supplies and know how to manage your diabetes without the loop.
Is there an iPhone version?
Yes, it is called Loop. It is a separate project with the same principles but different code and a different build process.
How do I get help if something goes wrong?
The community is active on Facebook, Discord, and the AndroidAPS Google group. But you need to understand the system yourself — no one else can troubleshoot your insulin settings for you.
Community builds
No community builds yet — be the first, we feature the best ones.
Discussion1
FROM THE COMPAREE TEAM
Thousands of people are running this right now, and the community has years of real-world tuning data. If you are already managing type 1 diabetes, what would tighten your control the most: overnight basal adjustments, meal predictions, or just never forgetting a correction bolus?
#WeAreNotWaiting community
The OpenAPS project was started by patients and parents who reverse-engineered the radio protocols of older insulin pumps and published everything so others could build the same system. AndroidAPS, Loop, RileyLink, and Nightscout are all community-led projects under the same banner.
Keep building — related projects
Same category, picked by how close they are to this build.
FULL CADA 3D printed stethoscope that costs 3 dollars matched a 200 dollar Littmann in a peer-reviewed test
HealthOpen-hardware●●○○○starter
FULL CADAn open-source hearing aid you build yourself and program in your own ear
HealthOpen-hardware●●●●○advanced
FULL CADYou can build the vital signs monitor from an ICU bedside - it runs on a Raspberry Pi
HealthOpen-hardware●●●●○advanced
FULL CADA mouse you control with your lips, for people who cannot use their hands
Open-hardwareHealth●●●○○moderate
DISCLAIMER
- Comparee is not the author of the projects featured here. All rights to each project belong to its creator — every page links to the original source, and we never host creators’ files.
- Information is provided without warranty and may become outdated as projects evolve. Prices are indicative bands only — always check the creator’s parts list for current costs.
- Building and operating any project is at your own responsibility. Protective equipment, safe workshop practice and compliance with local regulations are the builder’s responsibility.
CompareeTEAM13d ago
Practical notes from our verification: the payload links to the AndroidAPS repository, which is the Android app that closes the loop. The hardware rig (OpenAPS reference design) is a separate repository with full build docs, Gerber files for RileyLink boards, and 3D printable cases. There is no single official video — the best walkthroughs are community-made. The biggest surprise is that the pump compatibility bottleneck is real: the Medtronic models that work are discontinued, and finding one with the right firmware is expensive and time-consuming. The second surprise is how much of the difficulty is not the build, but the weeks of tuning with your clinical team afterward. This is a patient-built system with serious documentation and a large active community, but it is not a product and it is not for everyone.