“Closed loop.” “AID.” “Hybrid closed-loop.” “Algorithm-driven insulin delivery.” The vocabulary has run ahead of the explanation, and most people coming to it for the first time are reading marketing copy.

Here’s the calm version, with no manufacturer name in it.

What it actually is

A hybrid closed-loop system has three components working together:

A continuous glucose monitor (CGM) measuring glucose every few minutes.
An insulin pump delivering insulin continuously through a small cannula.
An algorithm running on the pump, on a phone, or on a small controller that reads the CGM and adjusts the pump’s basal insulin delivery automatically.

“Hybrid” is doing real work in the name. The system handles basal automatically. The person still has to announce meals telling the system “I’m about to eat about this many carbs” so it can pre-bolus. The system also can’t yet handle exercise without the person flagging it.

What it is not

It is not an artificial pancreas in the full sense. The pancreas does many things; HCL automates one of them.
It is not “set and forget”. It requires meaningful active engagement.
It is not a cure for Type 1.
It does not eliminate hypos. It significantly reduces them, particularly overnight.
It is not the same as the DIY Looping community’s open-source systems. See our DIY page.

What the evidence broadly shows

For most users in published clinical trials and real-world studies, HCL improves time-in-range, reduces nocturnal hypos, reduces hypo anxiety, and reduces the cognitive load of routine management. It also tends to reduce, but not eliminate, the “diabetes overhead” the constant background calculation.

It does not magically fix poorly-set carb ratios, sluggish meal-time pre-bolus habits, or systems-level issues. What it does is make a well-set-up system more forgiving and a tired user less penalised by tiny errors.

The trade-offs nobody mentions on the marketing site

Two devices, one body. CGM and pump live somewhere on you. Skin real estate matters. Adhesive sensitivities matter.
Alarm density goes up. Two devices, two sets of alarms. Tuning becomes important.
Phone dependency. Some systems run the algorithm on your phone. That phone now matters more than it did.
Algorithm trust. You’re handing some control to software. Many people find that liberating. Some find it disorienting at first.
Carb counting still matters. The system is only as good as the meal announcement.
Backup planning. You still need a Plan B for when any of the three components fails.

Who tends to benefit most

People with frequent nocturnal hypos.
People with high hypo anxiety affecting daily life.
People with significant glucose variability and good engagement with diabetes self-care.
Parents of young children with Type 1, where overnight management is the heaviest load.
People who already use a pump and are comfortable with diabetes tech.

Who may want to wait or skip

People in active burnout who can’t engage with a new learning curve right now.
People with severe alarm-fatigue concerns who haven’t tuned their current system.
People with skin or adhesive issues that already make sites difficult.
People without reliable access to the funding pathway or backup supplies.

Questions to ask your team

Of the systems available to me, which has the most independent evidence in someone like me?
What’s the funding pathway in my situation?
What does the onboarding training look like, and what’s the realistic learning curve?
What’s the system’s behaviour during exercise and unannounced meals?
What’s the failure mode and the backup plan?
How does the system handle me being unwell?
Can my diabetes team support this specific system long-term?

HCL is meaningful technology, used well. Used badly for the wrong reasons, at the wrong time, without proper onboarding it’s just more devices and more alarms. The framework helps you tell which one you’d be doing.

Related: Choosing a CGM