Comparison

Transkriptor versus the phone on the table: which job falls to which

Updated 14 September 2026 · Türkçe

Search for Turkish transcription and the first name that comes up is usually Transkriptor, and there is a good reason for that. But "turning an audio file into text" and "taking the notes for a meeting" are not the same job. This page separates the two.

What Transkriptor does

Transkriptor is a transcription service founded in Turkey and one of the more mature products in the category. As their own page puts it: speech-to-text in more than 100 languages, speaker recognition, file upload, a bot for online meetings (Zoom, Teams, Meet), a browser extension and iOS/Android apps.

So if you have an audio or video file — an interview archive, a YouTube video, a webinar recording, a podcast — Transkriptor is the right tool. The broad language support, the desktop side and the team plans are things we do not have.

Side by side

Prices and features taken from transkriptor.com and the App Store on 14 September 2026.

The real difference: does it sell minutes, or meetings

Transkriptor's plans are built on minutes — 300 minutes, 2,400 minutes. That is the right model for archive work: if you have 40 hours of interviews, you want to know how much you are consuming.

For someone who goes into three meetings a week, counting minutes is a strange exercise. The meeting is happening anyway; the point is not having to remember, record, upload and then read the result every single time. Ses Notu has no minute quota; the subscription opens the recording itself.

Why the speaker separation is not the same thing

Both say "speaker separation", but technically they are different jobs:

• Per-file separation. The service clusters the voices inside the recording you uploaded and labels them Speaker 1, 2, 3. It works correctly, but the labels belong to that file. In next week's meeting with the same team, "Speaker 2" may be someone else entirely.
• A persistent voice print. Ses Notu extracts a 192-dimensional voice print for each speaker and keeps it on the phone. The same person is recognised in the next meeting; name them once and the name sticks. The values we measured: two sentences from the same person score 0.83 similarity, different people 0.20.

That model (CAM++, WeSpeaker, Apache-2.0) ships inside the app and runs on the phone — because a voice print is biometric data and there is a good reason not to send it to a server. How it was done is described in a separate post.

Which one should you buy

Transkriptor, if:

• You have files to be transcribed — interviews, video, podcasts, webinars.
• You need languages other than Turkish and English.
• You work from Windows, Android or a browser.
• Putting a bot into your online meetings is not a problem for you — in fact it is what you want.
• You need a team plan, a shared workspace and per-seat billing.

Ses Notu, if:

• Your meetings happen in the same room — a client meeting, a team meeting, an interview, a class, a consultation.
• What you want is not a transcript but a meeting note: the decisions, and who owes what.
• You meet regularly with the same people and want speakers recognised from one recording to the next.
• You use an iPhone and prefer paying in lira.
• It matters to you that the recordings stay on the phone and are not uploaded to a cloud.

Using both is reasonable too: the meeting at the table to Ses Notu, the files in the archive to Transkriptor.