The Glucose Never Lies®
Science and lived experience
for type 1 diabetes
An independent education platform that translates complex diabetes science into clear, practical understanding. Built by clinicians and people living with type 1 diabetes, grounded in peer-reviewed evidence, and designed to help you explore how glucose really works.
How does Grace work?
Ask Grace anything about type 1 diabetes: medical devices, exercise, high-fat and high-protein meals, everyday life. She pulls the research together and answers at the reading level you choose, which you can change as your knowledge grows. For what to do for you or your child, ask your diabetes team; in an emergency, call 999.
Pick a tab to see one row on its own.
Answers at your reading level
In Ask Grace, pick how you like to read and she will answer that way; any level is open to anyone, and she starts at Advancing.
Level 1, Easy
5+ (read with a grown-up)Young children and families, newly diagnosed
Level 2, Straightforward
8+Living with diabetes
Level 3, Advancing
12+Experienced users, educators
Starts hereLevel 4, Clinical
AdultNurses, dietitians, doctors
Level 5, Scientific
AdultResearchers
What you can do with Grace
Ask and Explore are open to everyone; Build, Write and Review need an access token.
Open: Ask
About the evidence: a study, a claim, a device. In Ask Grace.
Open: Explore
GNL's educational explorers, from population evidence. Their modelling is GNL's own, so it sits at E on the A to E key: for learning, not for one person's dose. In Tools.
Needs a token: Build
Your own explorers, sites and platforms.
Needs a token: Write
Manuscripts and documents.
Needs a token: Review
Your work checked, so it stays honest.
Grace can help you build your own explorers, sites and platforms, write manuscripts, and review work so it stays honest. These need an access token. Contact john@theglucoseneverlies.com.
How the evidence is checked
When a paper comes into Grace, it is read closely and asked the same strict questions, in this order.
Question 1: How was the study built?
Each paper gets one letter for how the study was built, not how strong its advice is. Randomised means chance decides who gets what.
Question 2: What did they measure, and why?
How directly did the study measure a harm or benefit the person actually has, rather than a stand-in for one?
Question 3: Did they measure it well?
Was it measured for long enough to see a change? And often enough to see the path, and to know the change held?
Question 4: Were the results read right?
Did the authors read their own numbers fairly? Or were the results not stated?
Question 5: So what, for people with diabetes?
A judgement, with its reasons, on how useful the research is to people living with diabetes.
The A to E letter, for how a study was built
- AThe strongest randomised trials and reviews
- BOther cause-and-effect studies
- CReal-world observation
- DConsensus and guidelinesA guideline sits at D because the letter records how a source was built, not how strong its advice is.
- EUnvalidated, including GNL's own modelling
Created by two people with skin in the game
For people living with diabetes. Everything above stands on this.
Grace was developed around the papers supporting the ISPAD, ADA and EASD diabetes guidelines. She is built on around 1,000 of them. From that base, she takes in only papers that add value for people with diabetes, always asking: so what, for people with diabetes?
The Glucose Never Lies began with John's son, who tested positive for type 1 antibodies on a blood spot test. John needed a platform that could teach anyone who looked after his son. Grace draws on his 20 years of expertise, his publications and his skill in putting the evidence into practice at Birmingham Children's Hospital.
Phillip Hayes, John's best mate, is an expert in AI product development at a successful start-up in this field. He knows how to deliver the quality, safety and remarkable flexibility of an AI system, at a cost that keeps Grace free for people with type 1 diabetes.
For more on Grace's architecture, safety guardrails, ingest procedure, and medical and legal compliance, contact john@theglucoseneverlies.com.
Where to start
Start with Grace, then go deeper.
Three ways to get started
Grace is our AI education assistant for type 1 diabetes. Pick one of five reading levels and Grace answers questions at your level.
Ask Grace about the evidence. Explore her tools on AID systems, alcohol, exercise, and hypo and hyper management. Also included are a blood glucose meter selector and Survive and Thrive guides for CGM and AID systems.
Twelve essential building blocks: insulin, carbohydrate, activity, CGM, and glucose patterns.
Every Foundations module ends with an assessment, and a score of 9 out of 10 earns a personalised certificate.
Complete the Foundations to be ready for the advanced guides and Podcast.
Go deeper. In-depth guides on IOB, CGM technology, AID systems, and exercise physiology.
Every guide ends with a quiz, and a score of 9 out of 10 earns a personalised certificate.
The GNL Podcast
Latest episodes.
Tandem Control-IQ+ and Mobi, with Laurel Messer
The AID series opens: correction factor as the lever, what Control-IQ+ changes, and the matchbox sized Tandem Mobi.
Listen to Episode 43DICE 2026 Highlights
Highlights and key takeaways from the DICE 2026 conference.
Listen to Episode 42CGM Series Review and the Diabetech Conversation
Closing the CGM series, with a wide-ranging conversation on where diabetes technology is heading.
Listen to Episode 41Get in touch
Contact and follow.
Email Phill
AI solutions, including the technology that powers Grace.
phill@theglucoseneverlies.com