Comparison

Otter.ai does not support Turkish. What is used instead in Turkey?

Updated 14 September 2026 · Türkçe

Otter.ai is the best known name in the category. It is also of no use to you if you want a Turkish meeting written down — and that is not an accuracy problem, it is Turkish never appearing on the list of supported languages.

The fact first: Otter supports six languages

The wording on Otter's own pricing page: "AI transcription in English, Spanish, French, German, Japanese, and Chinese." Their help centre lists the same six. Turkish is not there. (Checked via otter.ai/pricing on 14 September 2026.)

Nor is the gap new. The Turkish technology press first wrote about it in 2020; six years later nothing has changed.

Not why Otter "struggles" with Turkish, but why it does not do it

Speech recognition models are trained per language. A model trained for English does not do Turkish "a bit badly"; it does not know Turkish phonemes, suffixes or word boundaries at all. Turkish is agglutinative — gelemeyeceklerini ("that they will not be able to come") is a single word, and a decoder trained for English cannot take it apart.

So "Otter is weak in Turkish" would be the wrong sentence. Otter does not do Turkish. That is a decision about which markets it sells to, not about the quality of the app.

The options that genuinely work for Turkish

The table below separates the tools that transcribe Turkish meeting audio and are reachable from Turkey by what they are for. They all do different jobs; which one is right depends on the shape of the meeting in front of you.

Prices taken from the providers' own pages on 14 September 2026. The only lira price in the table is Ses Notu's; the others mean paying in dollars from Turkey.

Four things to check when choosing

1. Is Turkish inside the model, or only in the interface? A tool with a Turkish website does not mean it can decode Turkish speech. Speak a few sentences into a trial recording and look.
2. Does speaker separation hold up in Turkish? Diarization works independently of language — it looks at the voice print, not the word. But if the tool only does it in an online meeting, from a separate audio channel per participant, it will not help you when you record in one room.
3. Where does the meeting happen? If your team is around one table, every bot-based tool is out: there is no link to join.
4. Payment. A monthly dollar subscription comes with the exchange rate attached, and paying it on a company card is a job of its own. A lira subscription through the Turkish App Store is also less work on the accounting side.

Where Ses Notu sits in this list

Ses Notu is not a cheaper version of the above; it was written for a different meeting. The meeting where everyone is in the same room. You put the phone in the middle of the table and press record; speech turns into text as people talk, speakers are separated from one another, and when the meeting ends you get a summary, the decisions taken, and the action items with an owner and a deadline.

Two technical details matter for Turkish:

• Speaker separation runs inside the phone. The model that extracts the voice prints (CAM++, ONNX Runtime, ~28 MB) ships inside the app. A voice print is biometric data; it never leaves the phone. The detail is in this post.
• The output is a Turkish meeting note. Not just a transcript: "by Friday" is turned into a real date, and action items go to Reminders with one tap.

To be straight about it: Ses Notu only runs on iPhone and Apple Watch. There is no Android or desktop version. If you have years of accumulated audio and want to transcribe it in bulk, this is not what you are looking for — the Transkriptor comparison sets out which job falls to which tool.