Speaker identification
Kuno automatically detects who's speaking in every recording and separates their contributions in the transcript. You assign a name to each speaker once — Kuno then recognizes them automatically going forward, and all voice profiles stay EU-hosted and confined to your account alone.
At a glance
- Kuno separates speakers automatically and initially labels them "Speaker 1," "Speaker 2," and so on in the transcript
- You name speakers once in the Kuno app — the assignment then applies to future recordings
- Incorrectly split voices can be merged, renamed, or removed
- Voice profiles are biometric data (GDPR Article 9) and are hosted exclusively in the EU, encrypted, and processed only within your account
- Kuno does not use your voice profiles to train AI models
- Deleting a speaker profile keeps the recording intact — only the name assignment is removed
Speaker identification assigns each contribution in a recording to the right person. This keeps even long conversations easy to follow, and lets summaries, decisions, and action items be attributed to the correct person.
How Kuno identifies speakers
As soon as you finish a recording on your Kuno device, the Kuno app analyzes the audio and separates the individual voices. At first, they appear neutrally as “Speaker 1,” “Speaker 2,” and so on. Identification runs entirely on EU servers — your recordings never leave the European legal area.
Naming speakers
- Open the recording in the Kuno app.
- Tap a speaker label in the transcript (e.g. “Speaker 2”).
- Enter the person’s name and confirm.
Kuno creates a voice profile from this. If Kuno recognizes the same voice in a later recording, the name is assigned automatically — no need to name them again.
Merging or correcting speakers
If one person is accidentally split into two speakers — for example due to call quality or background noise — you can merge the profiles:
- Open Settings › Speakers in the Kuno app.
- Select the affected profiles.
- Tap Merge.
You can also rename speakers or remove a profile here.
Note on privacy: Under GDPR (Article 9), voice profiles are considered biometric data and therefore especially sensitive. Kuno stores them encrypted, hosted exclusively in the EU, and processed only within your account. They are never shared, used for cross-context identification, or used to train AI models. Also keep in mind: anyone being recorded must consent (Section 201 StGB). Get the consent of everyone involved before naming and permanently assigning voices.
Deleting a speaker profile
Deleting a speaker profile under Settings › Speakers only removes the name assignment. The recording and transcript remain intact — the person in question then appears again as an anonymous “Speaker.” This way, you retain full control over the stored biometric data at all times.
Our team is happy to answer questions about speaker identification at [email protected].