vastflow
Abstract
We introduce vastflow, an independent studio that writes, publishes, and maintains its own mobile applications for iOS and Android. In this short note we outline the principles that inform our practice, supply a technical card describing our output, and provide an address for correspondence. We report no findings. We make no predictions about our field. We simply write the software, and then we return to it1.
1Introduction
Most software is written by committees. The result, predictably, reads like committee minutes. We propose an alternative: a studio of limited size whose members write all of their own code, make all of their own mistakes, and are therefore obliged to live with both2. vastflow is the name we have given to this arrangement.
The studio takes its name from the observation that the field of mobile software is, on inspection, less a torrent of new ideas than a vast flow of old ones, reasserted with small local variations. Our contribution, such as it is, is not to the flow, but to what arrives at its banks, is picked up, is used, and — if we are careful — continues to be useful tomorrow. At time of writing, the studio's principal application is SozAI, an AI transcription tool that turns audio, video, and YouTube into searchable text across more than a hundred languages.
2Practice
Our method is unremarkable. We identify a problem that we ourselves have, or that we can understand through patient observation; we design an application that addresses it; we ship the application; and we continue to maintain it, without limit, until we are no longer able to3. We work in Swift, Kotlin, and Dart. We read platform guidelines and follow them, because it is rude not to.
Publishing our own work also means buying the advertising that brings people to it, and measuring whether that was money well spent. We do the first in Apple Search Ads and Google Ads, and the second with a private dashboard of our own; both are described in § 2.4, Instrumentation.
Equation (1) is not meant seriously as a measurement4. It is meant as a commitment: we believe that no amount of care at release compensates for its absence afterward, and that the only honest way to describe an application's quality is over the full arc of its life.
3Model Card
A brief specification of the studio's output, in the manner of a dataset card. Values are approximate and subject to drift.
| Domain | Mobile software |
|---|---|
| Platforms | iOS ≥ 17 · Android ≥ 13 |
| Languages | Swift 5.x · Kotlin 2.x · Dart 3.x · SwiftUI · Compose · Flutter |
| Team size | small |
| Output | Self-published applications |
| Distribution | Apple App Store · Google Play |
| Acquisition | Apple Search Ads · Google Ads — our own accounts, bought at our own expense |
| Clients | None. We do not take on commissioned work. |
| Release cadence | Irregular, by necessity |
| Maintenance | Indefinite |
| License | Proprietary (application), MIT (library code, where applicable) |
| Contact | [email protected] |
| Status | ● active — maintained |
4Applications
A representative selection of what the studio publishes. Each is available on the App Store, several also on Google Play, and each keeps a site of its own where it is described at length.
| Nomax | Travel. Puts solo travellers who are in the same city on the same evening in touch with one another. |
|---|---|
| SozAI | Transcription. Audio, video and YouTube turned into searchable text across more than a hundred languages. |
| notavik | Clinical notes. Records a consultation, separates the voices, and drafts a structured note — on the device, without uploading the audio. |
| Telvio | International calling over the internet, paid by the minute, without a subscription. |
| Orvik | Identification by camera: plants, birds, insects, rocks, coins and other small things one wants named. |
5Correspondence
We read correspondence at the address [email protected] and reply, as a rule, within one working day. Letters concerning defects in our applications are appreciated; letters proposing features are appreciated also; letters asking only what we think about the industry are answered but, we warn, briefly. For anatomy of a useful letter, see § A, Correspondence.
References
- Knuth, D. E. Literate Programming. The Computer Journal, 27(2), 97–111, 1984.
- Dijkstra, E. W. The Humble Programmer. Communications of the ACM, 15(10), 859–866, 1972.
- Gall, J. Systemantics: How Systems Work and Especially How They Fail. Quadrangle, 1975.
- Apple Inc. Human Interface Guidelines. developer.apple.com, 2026.
- Google. Material Design 3. m3.material.io, 2026.
- Hunt, A. & Thomas, D. The Pragmatic Programmer (20th anniversary ed.). Addison-Wesley, 2019.
cite as@misc{vastflow2026, author = {vastflow Research}, title = {vastflow: A Small Studio for Mobile Software}, year = {2026}, note = {Almaty, KZ · https://vastflow.kz}, }