DreamSync

Sleep apps have a credibility problem, and it comes from two directions at once. They measure something they cannot measure directly, and they rarely name any method behind their advice.

A phone on a mattress does not read your brain. It infers sleep stages from movement and sound, which is a reasonable proxy and not the same thing as the polysomnography those stage names come from.

DreamSync is more interesting than most of the category because it addresses the second problem squarely: it names CBT-I, which is a specific intervention with a large clinical literature behind it, rather than offering unattributed advice. That is worth taking seriously, and so is the equally unusual decision to keep every byte on your phone.

What Does DreamSync Do Differently?

• It runs a 12-week CBT-I programme — cognitive behavioural therapy for insomnia, the intervention with the strongest evidence base for chronic sleep problems.
• Everything stays on your device: no servers, no account, no sign-up, no tracking and no ads, with data in files you can inspect, export or delete.
• Behaviour correlation is the analytical core — how caffeine, exercise, screen time, stress and alcohol actually affect your sleep, based on your own logged data.
• Tracking with sleep-stage analysis, a smart alarm, sleep score, efficiency tracking, and snoring and disturbance detection.
• Works completely offline, with data export for healthcare providers and trends across 7, 30 and 90 days.

Why Naming CBT-I Changes The Conversation

DreamSync includes a 12-week clinical CBT-I programme. Cognitive behavioural therapy for insomnia is not a wellness framework — it is a structured intervention with a substantial evidence base, and major clinical guidelines recommend it as first-line treatment for chronic insomnia, ahead of medication.

That matters because almost nothing else in this category names a method. Most sleep apps offer tracking plus general advice — go to bed earlier, avoid screens — which is unobjectionable and does very little for someone with genuine insomnia.

CBT-I is different because it is counterintuitive in ways that general advice never is. Its core components include sleep restriction, which temporarily reduces time in bed to consolidate sleep, and stimulus control, which means getting out of bed when you cannot sleep. Both feel wrong, and both are what makes the intervention effective.

A twelve-week structure is also the right shape. CBT-I works over weeks rather than days, and the early stage frequently makes things feel worse before they improve — which is precisely why an app promising better sleep tonight cannot deliver this.

The honest caveat is that self-guided CBT-I is not the same as clinician-delivered CBT-I. The evidence for digital versions is real and generally weaker than for supervised programmes, and someone who cannot get through the difficult early weeks alone may need actual support. If insomnia is severe or persistent, see a doctor — an app is a reasonable first step, not a substitute for one.

It is worth being clear about the boundary between this and a meditation practice. Apps like StillMind and Insight Timer help you wind down, which is genuinely useful and is not an insomnia intervention.

CBT-I works partly because it is counterintuitive. An app promising better sleep tonight cannot deliver an intervention whose first weeks make things harder.

What Phone-Based Sleep Staging Can And Cannot Tell You

DreamSync offers sleep-stage analysis, a sleep score and efficiency tracking, and the limits of phone-based measurement deserve stating plainly.

Sleep stages are defined by brain activity, measured in a laboratory with electrodes. A phone infers them from movement and sound, which correlates with sleep depth well enough to be interesting and not well enough to be diagnostic.

The practical consequence is about which numbers to trust. Total sleep time and sleep efficiency — time asleep as a proportion of time in bed — are reasonably robust from movement data, and efficiency is also the number CBT-I actually uses. Stage breakdowns are much softer, and a chart telling you precisely how much deep sleep you had is more confident than the method supports.

Snoring and disturbance detection sits somewhere useful in between. A phone genuinely can record sound, and knowing you snore heavily or wake repeatedly is real information — occasionally important information, since loud snoring with pauses in breathing is a reason to see a doctor about sleep apnoea rather than to adjust your bedtime.

Read the trend rather than the night. A single night's figures are noise; a pattern across ninety days is the signal these tools genuinely provide.

How Serious Is A No-Servers Privacy Position?

DreamSync stores all sleep data locally on the device in files you can inspect, export or delete, with no account, no tracking and no ads, and works completely offline.

That is an unusually strong position and it deserves crediting, because of what sleep data actually contains. Combined with the behaviour logging this app encourages, the dataset covers when you sleep, how badly, how much you drink, how stressed you are and how much caffeine you consume.

That is health data by any reasonable definition, and it is exactly the kind of information that has commercial value to insurers, employers and advertisers. Most consumer health apps hold it on a server under a privacy policy that can change.

Local-only storage removes the question entirely rather than answering it. There is no server to breach, no policy to change, and no acquisition that transfers your data to someone whose practices you never agreed to.

Two practical consequences follow. Your data is only as safe as your phone and your backups — losing the device means losing the history, so use the export. And no cloud means no sync, so this is a single-device product by design. Both are reasonable trades for what you get.

• Export periodically, since local-only means a lost phone is a lost history.
• Expect no sync across devices; this is single-device by design.
• Use the healthcare export if you are seeing a doctor about sleep — real data beats recollection.
• Log behaviours honestly, including alcohol, since the correlations are only as good as the inputs.
• Treat loud snoring with breathing pauses as a reason to see a doctor, not to adjust your bedtime.

Behaviour Correlation Is The Genuinely Useful Output

DreamSync logs caffeine, exercise, screen time, stress and alcohol, and correlates them with your own sleep data.

This is the most valuable thing it does, and the reason is that general sleep advice is true on average and unhelpfully so. Everyone knows caffeine affects sleep; almost nobody knows whether their own afternoon coffee matters, and the answer varies enormously between individuals for genuine metabolic reasons.

Personal correlation answers the question that generic advice cannot. It may turn out that caffeine has little effect on you and alcohol has an enormous one, or that the variable driving your bad nights is stress rather than anything you consume.

Alcohol is the one most people are wrong about, and the correlation usually shows it. A drink helps you fall asleep and reliably degrades the second half of the night, which is why people who use it as a sleep aid wake at four and cannot understand why.

The caution is statistical rather than technical. Correlation across a handful of nights is noise, and the app cannot separate causes that move together — a stressful week is also a week you drank more and exercised less. Treat a pattern as a hypothesis to test deliberately, by changing one thing for a fortnight, rather than as a finding.

The same logic underlies food and mood tracking. Tools such as MyNetDiary exist because population averages tell you very little about your own body, and sleep is no different.

Running The Twelve Weeks Properly

A CBT-I programme only works if completed, and the first fortnight is where most people stop. That is a feature of the intervention rather than a flaw in the app.

The bar is step four. If you abandon during the hard weeks, you have tested nothing about whether CBT-I works for you.

1. Track for two weeks before starting the programme, so you have a genuine baseline to measure against.
2. Log behaviours honestly from the start — especially alcohol, which is the variable people most often under-record.
3. Read what sleep restriction and stimulus control actually involve before beginning, since both feel wrong and knowing why helps you continue.
4. Expect the first two to three weeks to feel worse rather than better. That is how the intervention works, and it is where people quit.
5. Watch sleep efficiency rather than stage breakdowns, since efficiency is both more reliably measured and what CBT-I actually targets.
6. Export your data before the twelve weeks end, both as a backup and to bring to a doctor if things have not improved.

Where An App Stops And A Doctor Starts

This deserves stating plainly rather than as a disclaimer, because sleep problems are frequently symptoms of something an app cannot address.

Loud snoring with pauses in breathing, or waking gasping, may indicate sleep apnoea — a condition with serious cardiovascular consequences that requires diagnosis and treatment. A sleep app that detects snoring is genuinely useful for surfacing this, and completely unable to treat it.

Persistent insomnia can also be secondary to depression, anxiety, thyroid problems, chronic pain or medication side effects. Treating the sleep without addressing the cause does not work, and the pattern of not improving despite doing everything right is itself information worth taking to a doctor.

The healthcare export is the feature that bridges this properly, and it is worth using. A clinician asking how you have been sleeping gets a far better answer from ninety days of data than from anyone's recollection, which is reliably poor.

None of this argues against the app. Self-guided CBT-I is a reasonable first step for ordinary insomnia, and it is far better than the alternative most people choose, which is nothing. It is a first step rather than the only one.

If low mood rather than sleep is the underlying problem, a support tool like Youper addresses a different part of the same picture. Neither replaces a clinician.

DreamSync: Method And Data Handling

Who Should Try DreamSync

A sensible first step when

• You have ongoing insomnia and want an intervention with an evidence base rather than general advice.
• You are unwilling to put sleep, alcohol and stress data on someone else's servers.
• You want to know how your own caffeine and alcohol affect your sleep rather than what affects people on average.
• You will see a doctor about sleep and want real data rather than recollection to bring.
• You can commit twelve weeks, including two or three that feel worse before they improve.

Go elsewhere when

• You snore loudly with pauses in breathing — see a doctor about sleep apnoea rather than starting a programme.
• Your sleep problem is likely secondary to depression, anxiety, pain or medication, where treating sleep alone will not work.
• You are on Android, since this is an iOS application requiring iOS 15.1 or later.
• You need data synced across devices, which local-only storage rules out by design.
• You want better sleep tonight, which no CBT-I programme promises or delivers.