Voice Capture
Catches the thought before it gets away — and turns it into a transcript, notes and tasks. Private, in Europe.
Voice recorder & transcript and AI summary
Coming to the App Store beta test in progressNo sign-up. You start in ten seconds.
Illustrative visuals created with artificial intelligence. The app screenshots are from the real app.
Voice Capture is a voice notebook for people whose thoughts run faster than they can write them down. No sign-in, no meeting setup — you just record.
All of it works with one thumb, even walking and without looking — with haptic feedback.
Home Screen widget, Action button and Lock Screen — recording is always one touch away.
The transcript is an immutable original — never overwritten, never lost, never quietly “improved”. On top of it you pick a template and AI produces a separate output. Edit outputs or create them again; the original stays untouched.
Advanced analysis hears individual speakers, tone and ambient sound. One meeting gives you notes for yourself and a report for the client.
None of the ready-made ones fit? With Premium you write your own: describe in your own words what AI should make from a recording, and the template shows up in the menu next to the others. You can keep up to five of them and rewrite them anytime.
Recordings laid out by day, with title, summary and length at a glance.
On newer iPhones the transcript is created on the device itself. You decide what goes to the servers and what AI may process — permanently in Settings, or just for a single recording.
See exactly where your data sits →
Settings draw the data path for you — and the nodes your data never reaches are greyed out. The picture shows your setup, not a generic diagram.
The web adds a “Where my data is” overview: exactly what sits on the servers. And you can delete all of it, account included, whenever you want.
Transcript to the servers, audio to the servers, AI processing — you choose each one separately. The recording screen shows what will happen, and a tap changes it for a single recording.
On-device transcription needs iOS 26 and a downloaded language model. On older iPhones the transcription runs on servers in the EU.
Transcription, summaries, speaker labels, the app and support — all built for Czech from the start, not machine-translated from English. The same care goes into English.
Speaker separation (“Speaker 1”, “Speaker 2”) is included, not an add-on you pay extra for.
What is yours stays yours. No vendor lock-in.
Recordings, transcripts and every AI output are waiting in your browser at voice-recorder-2026.web.app/moje — with the spoken word highlighted during playback. Just sign in with Apple ID.
With a personal key, anything else can read your finished transcripts — Notion, Zapier, Shortcuts, your own script or your journaling app. You create the key in the app with one tap and revoke it whenever you want.
The key can only read: no uploading, no deleting, no touching your account. You can take your data out by export as well; nothing locks you in.
Three steps. The first one is the only one you do.
Tap or hold the button and talk. Nothing else to set up.
The transcript is created right on the phone and becomes an immutable original.
AI turns the transcript into notes, tasks or a journal — as many times as you like.
Voice Capture is in beta and heading to the App Store.
Coming to the App Store beta test in progress