Cognitive behavioral therapy for insomnia — CBT-I — is the treatment every sleep guideline in the world recommends before any pill. It works. It works durably. And for about three decades it has been almost impossible to actually get, because the number of clinicians trained in it is a tiny fraction of the number of people with chronic insomnia.
Digital CBT-I was built to close that gap. Deliver the same protocol — sleep restriction, stimulus control, cognitive restructuring, sleep hygiene, relapse prevention — through a phone or web app, and reach a hundred times more people at a fraction of the cost. The evidence says it mostly works. The market says it is a mess. Big products have folded, prescription pathways are inconsistent, and a large slice of what gets marketed as a sleep app has nothing to do with CBT-I at all.
This is the honest comparison. What is CBT-I actually. What does the evidence show. Which apps deliver the real thing. Which ones are meditation apps in a sleep skin. And the free options nobody talks about.
What CBT-I actually is
CBT-I is a structured, multi-week protocol. It has five load-bearing components.
- Sleep restriction. Compress your time in bed to match your actual sleep time, then slowly expand it. The point is to consolidate broken sleep into a single solid block. This is the single most powerful lever in the protocol and the one people hate the most.
- Stimulus control. Bed is for sleep and sex only. If you cannot sleep within about 20 minutes, get up. Rebuild the mental link between the bed and sleep onset.
- Cognitive restructuring. Identify and challenge catastrophic thoughts about sleep. "If I do not sleep tonight my life is over tomorrow" is the specific cognition CBT-I targets.
- Sleep hygiene. Cool, dark, quiet, no caffeine after noon, consistent bedtime. Necessary but never sufficient.
- Relapse prevention. A written plan for when insomnia comes back, because it does.
An app that delivers only meditation, only sleep sounds, or only tracking is not delivering CBT-I. Keep that distinction in your head for the rest of this article.
The evidence base
Digital CBT-I has one of the strongest evidence bases in consumer health tech.
The Trauer 2015 meta-analysis in Annals of Internal Medicine pulled 20 randomized trials of face-to-face CBT-I and found effect sizes of roughly 0.6 to 0.8 standard deviations on sleep efficiency, wake-after-sleep-onset, and sleep latency. Those are large effects for a behavioral intervention. Subsequent meta-analyses of the digital versions — Zachariae 2016 and Soh 2020 — landed at somewhat smaller but still robust effects, generally in the 0.4 to 0.6 range across the same outcomes.
The Ritterband 2017 SHUTi randomized trial, published in JAMA Psychiatry, followed 303 adults through fully automated internet CBT-I and found sustained improvements in insomnia severity at one-year follow-up. That durability matters. Sleeping pills stop working when you stop taking them. CBT-I keeps working after the treatment ends.
The Espie 2012 randomized trial of Sleepio, published in Sleep, found roughly a 20 percent absolute improvement in sleep efficiency in the treated group versus control, with effects maintained at eight weeks. NICE — the UK's National Institute for Health and Care Excellence — reviewed the evidence and recommended Sleepio in 2022 as a first-line treatment for adults with insomnia.
Two caveats matter as much as the effect sizes.
Adherence. Digital CBT-I has dropout rates around 30 to 40 percent across most trials. The apps that produce the biggest effects are also the ones users are most likely to abandon, because sleep restriction is genuinely unpleasant for the first two weeks. Any comparison of apps that ignores adherence is misleading.
Selection bias. Trial populations tend to be motivated volunteers. Real-world effect sizes are typically smaller than trial effect sizes. Expect a good, not miraculous, result.
Sleepio (Big Health)
Sleepio is the most-studied digital CBT-I product in the world. Big Health, the UK company behind it, has published more than a dozen randomized trials. The Espie 2012 trial is the canonical evidence. NICE approval in 2022 makes it the first digital therapeutic recommended as a first-line insomnia treatment inside the UK's National Health Service.
The product itself is a six-session animated program guided by a virtual therapist character called "The Prof." You complete a sleep diary daily and the algorithm adjusts your sleep window based on the data. The core CBT-I protocol is there — sleep restriction, stimulus control, cognitive work, and relapse prevention are all delivered in the intended order.
In the US the picture is more complicated. Sleepio is available through employer health plans and some direct-to-consumer channels at roughly $400 per year retail, though most users access it through a covered benefit. If your employer or insurer offers it, take it. It is the closest thing to a proven, credentialed digital CBT-I product currently in market.
Somryst (Pear Therapeutics)
Somryst was FDA-cleared in 2020 as the first prescription digital therapeutic for chronic insomnia. The underlying protocol is the SHUTi content adapted for prescription delivery, so the Ritterband 2017 trial data underwrites the clinical evidence.
The complication is Pear Therapeutics filed for bankruptcy in April 2023. Somryst's US commercial status has been shaky ever since. As of writing, availability depends on which relaunched entity is currently servicing prescriptions. The underlying clinical content is genuine. The product's continuity is not something you can build a treatment plan around right now. If your clinician offers it stably, use it. Otherwise, choose something with a stable commercial path.
Stellar Sleep
Stellar Sleep is a newer direct-to-consumer app that markets itself as a CBT-I product for chronic insomnia. Positioning is "psychology-based, not sleep hygiene, not meditation," and the marketed protocol does implement CBT-I components. Pricing is subscription, roughly $250 to $350 for a full program.
The honest read: Stellar is executing the right pattern with light peer-reviewed evidence behind the specific product. Founders have published some efficacy data, but there is no equivalent to Espie 2012 or Ritterband 2017 for the Stellar product itself. That is a reasonable bet — the CBT-I protocol is well-defined and a competent implementation should produce the predicted effects — but it is not the same level of evidence as Sleepio. Price it accordingly.
CBT-i Coach (VA, free)
This is the app almost nobody outside veterans' health circles talks about, and it is the one you should probably try first.
CBT-i Coach was developed by the US Department of Veterans Affairs' National Center for PTSD in partnership with Stanford. It is free, on iOS and Android, and it is designed to be used either alone or alongside a clinician-led CBT-I program. It walks you through the same components as the paid products — sleep diary, sleep restriction calculations, stimulus control instructions, relaxation modules, and cognitive strategies.
The user experience is functional rather than polished. There is no animated therapist, no daily nudging, no motivational scaffolding. You have to bring the discipline. In exchange, you get an actual CBT-I toolkit, backed by an institution with deep clinical expertise, for zero dollars.
For a self-motivated adult with clear insomnia symptoms and no immediate access to a clinician, this is the honest starting point. It is not a coincidence that many CBT-I clinicians recommend it to their own patients as a companion to therapy.
Insomnia Coach (VA, free)
Insomnia Coach is a companion product from the same VA program, aimed at self-guided use rather than clinician-adjunct use. It is a five-week program with more structure and hand-holding than CBT-i Coach, delivered as a series of weekly modules with daily check-ins. Same evidence-based content, better on-ramp for someone starting solo.
If CBT-i Coach feels like a reference manual, Insomnia Coach feels like a workbook. Either is a legitimate free option. Try both, keep the one that fits your working style.
Calm, Headspace, Better Sleep — not CBT-I
This section exists because the confusion is chronic and expensive.
Calm, Headspace, Better Sleep, Balance, and the dozens of similar apps in the wellness category deliver meditation, breathwork, sleep stories, and guided relaxation. Some of that is useful. Some evidence supports meditation-based interventions for sleep in general populations, with effect sizes smaller than CBT-I. None of it is CBT-I.
These apps do not deliver sleep restriction. They do not walk you through stimulus control. They do not restructure catastrophic sleep cognitions. They will not compress your sleep window and expand it as your efficiency improves. If you have chronic insomnia and you are paying for Calm hoping it does what CBT-I does, you are paying for the wrong tool.
Meditation apps can complement CBT-I well. They can serve as a wind-down tool inside a stimulus control protocol. They are not a substitute for the protocol.
Objective data during sleep restriction
Sleep restriction is the most powerful component of CBT-I and the one where objective data helps most. Subjective sleep diaries are the standard input, and they work — but they are noisy, especially during the first two weeks when patients often catastrophize a bad night as "I didn't sleep at all" when they actually slept four fragmented hours.
A non-contact sleep tracker under the mattress can supply the objective anchor. Total sleep time, sleep efficiency, and wake-after-sleep-onset — the exact metrics that drive sleep restriction calculations — logged nightly with no wearable to remember and no cognitive load.
This is a supporting tool, not a treatment. The Withings mat by itself does not do CBT-I. Combined with CBT-i Coach or a paid digital CBT-I program, it makes the sleep restriction phase more accurate and less demoralizing.
The adherence problem
Every digital CBT-I product loses roughly 30 to 40 percent of users before completion. This is the number the marketing does not lead with, and it matters more than any brand comparison.
Three practical moves reduce your personal dropout risk.
- Start when your schedule is stable. Sleep restriction is brutal in the first two weeks. Do not start the week you begin a new job, travel across time zones, or have a newborn.
- Tell one person. A partner, a friend, a clinician. Adherence to any behavioral protocol roughly doubles when someone else knows you started.
- Commit to six weeks, not two. Meaningful improvement typically shows up between weeks three and six. If you quit at week two because you feel worse, you quit exactly when you were about to feel better.
A decision framework
Here is the honest order.
- Try CBT-i Coach or Insomnia Coach for four to six weeks. Free, evidence-based content, no barriers. If you complete it and your insomnia is resolved, you saved yourself a subscription.
- If your employer or insurer covers Sleepio, use it. Best-studied product, cleanest evidence, structured guidance that reduces dropout risk relative to fully self-guided options.
- If Somryst is stably available through your clinician, that is a legitimate prescription path. Verify current commercial status before you build a plan around it.
- Stellar Sleep is a reasonable direct-to-consumer paid option if the free apps do not fit your working style and covered access to Sleepio is not available. Understand that the specific-product evidence is thinner than Sleepio or Somryst.
- See a CBT-I clinician if the app-only route is not working after eight weeks. The Society of Behavioral Sleep Medicine maintains a public directory.
What is not on this list: Calm, Headspace, Better Sleep, Balance, or anything else that markets meditation and sleep stories as insomnia treatment. Those are useful for a different problem.
The bottom line
Digital CBT-I is real, and the evidence is stronger than almost anything else in the sleep-tech consumer space. Sleepio is the best-studied paid product. The two free VA apps deliver most of the same content for zero dollars and are chronically underused. Stellar Sleep is a plausible direct-to-consumer option with lighter product-specific evidence. Somryst's clinical content is genuine but its commercial status is unstable. Meditation apps are not CBT-I regardless of the marketing copy.
Pick the cheapest option that will keep you engaged for six weeks. Track your sleep with objective data if you can. Expect the first two weeks to feel worse before things improve. That is what the protocol working looks like.
