Reviews · Review
The Second Opinion: MindIsland, heard again after version 2.9
From 1.0 to 2.9 in barely three weeks: we ask whether such a pace of updates suits an app whose business is calm.
By The Editors · · 5 min read
Theatre critics know the ritual. A production has its premiere, the reviews appear, and weeks later one goes again to see what has settled. Some performances improve with time as the actors grow into their roles. Others lose the tension of the first night. The ‘second opinion’ is therefore a venerable form, and we apply it here to an app that is changing at breathtaking speed.
MindIsland, the AI meditation app, appeared in the App Store on 16 September 2026 in version 1.0. By 5 October 2026 it already bore the number 2.9. That is not even three weeks. For an app whose core promise is calm, that is a remarkable pace, and it raises a question we take seriously: does this pace serve calm?
What we know, and what we don’t
First, the honesty that a second opinion demands. We know the version numbers and the dates. We do not have a complete, publicly documented list of changes between 1.0 and 2.9. We will therefore not pretend to know which feature arrived in which version. Criticism that invents changelogs is not criticism but fan fiction.
What we can describe is the state of play shown by the website and store listings today. And it is remarkably complete for a product in its first month: personalised meditations from a text prompt, a fresh daily meditation, lengths of three, five or ten minutes, several AI voices, optional background music, a library in eight categories, an article collection with a daily quote, streak counter and history, offline storage of the last ten sessions, and data export and account deletion under GDPR.
The constant: the first breath
At a revival, one first notices what has stayed the same. With MindIsland, it is the opening. Every session begins, according to the maker, with the same breathing pattern: four seconds in, a short pause, six seconds out. Then follows the three-step sequence that distinguishes the app from many competitors: acknowledge the problem, work through and reframe it, then come to rest.
This structure is the backbone of the product. As long as it holds, much can be built around it without the experience falling apart. A theatre that constantly rebuilds the stage but does not touch the play can live with that.
A theatre that constantly rebuilds the stage but does not touch the play can live with that. It only becomes dangerous when the actors are given a different script every night.
The problem with the pace
And yet a discomfort remains. Meditation is a practice of repetition. One returns to the same place, to the same voice, to the same motions, and it is precisely this familiarity that creates the space in which something can come to rest. An app that changes weekly risks that listeners must first relearn the setup every time they open it.
We do not know how many of the updates brought visible changes and how many were invisible repairs. Young apps often receive bug fixes in quick succession, and rightly so. No one wants a meditation that crashes mid-exhale. But that is precisely why we would like the maker to name its changes clearly. A programme belongs with every revival.
What the speed reveals
One can read the pace more charitably. An app that patches so quickly evidently takes feedback seriously. A product whose content is generated anew at every moment is by nature more agile than a library of studio recordings. At Headspace or Calm, a new voice would be a production project. In an app that writes text via the Gemini interface and voices it using Inworld’s speech synthesis, it is, at a pinch, a setting.
This agility is both strength and weakness. It allows rapid improvement, but it also makes it hard to predict what one will hear tomorrow. For critics, that means our judgements about the voice or the wording of individual sessions have a shorter shelf life than with a classic meditation app.
What has stayed the same: open questions
Some points from our first review still stand unchanged. After the first free meditation, a subscription is required; the monthly subscription cost $12.99 in the US App Store and €14.99 in the German store on 10 October 2026. How many personal meditations are included in the subscription cannot be clearly deduced from the store listings, because quotas for additional meditations are sold at the same time. Meditations in English are documented; we cannot confirm other languages. And the app still explicitly does not describe itself as a medical device. That is as it should be, and we praise the clarity.
How we measure a revival
So that our next visit is fair, we set out our criteria. First, stability: a meditation that stutters, cuts out or fails to load misses its purpose entirely, and rapid updates are only a gain if they eliminate such faults. Second, consistency of form: the three-step sequence of acknowledging, reframing and resting, along with the opening breath, are the foundation. We will listen closely to see whether they remain. Third, transparency: we would like clear release notes, a clear statement of how many personal meditations the subscription includes, and an indication of the languages in which the meditations are generated.
Fourth and finally, the calm of the interface. A meditation app should feel familiar when opened. New features are welcome as long as they do not lengthen the path to the next session. A theatre may renovate its foyer, but it should not move the cloakroom to a different place every week.
Our second opinion
Our first encounter with MindIsland showed us that the idea holds: a meditation that first says what it is about feels more credible than any invitation to an imaginary beach. The second encounter confirms that idea and adds a concern. An app for calm must itself be a calm place, including in its development.
Those who would like to get to know the app can find it at mindisland.app. We will listen again in a few months, then hopefully with a detailed programme in hand.