How to create speaker profiles in HoldSay
Name a voice once for future recordings.
A speaker profile connects a recurring voice with the name you want to see in your notes.
Your own voice is learned automatically. Any other voice can be named once from a clear recording, and from then on real names replace generic Speaker labels in past and future notes.
Name a voice once — every future recording knows who is speaking.
Your voice is learned automatically
HoldSay learns your own voice without any setup. The test dictation you speak during first-run setup is saved locally as a profile labeled Me (shown in your interface language), so you are recognized by name from the very first recording. If that sample was too short, HoldSay quietly backfills your voiceprint from your recent push-to-talk dictations, since those are almost always you.
Recognition itself plays no favorites. Every voice, yours included, is compared against all saved voiceprints and named only when the match is confident; when HoldSay is unsure, it leaves a generic Speaker label rather than risk a wrong name. If you are ever missed, name yourself once in any note and HoldSay will recognize you across the library from then on.
Saved people appear with real names, photos and reusable local speaker profiles.
Your profile sits first in Settings under People, where you can set your real name, add a photo, or play the stored sample. When a match is especially confident, HoldSay also folds that recording's voiceprint into the profile, so recognition keeps improving across rooms and microphones.
Name a recurring voice
For everyone else, start from a recording where the person speaks clearly: a few uninterrupted sentences, little background noise, no overlapping talk. One clear passage beats many noisy ones.
Name the person from a finished recording and add a real photo when useful.
- Open the finished recording and switch to Speakers.
- Select the generic speaker label for the person you want to name.
- Enter their name, and add a photo if it helps you tell people apart.
The name applies to the whole library at once: HoldSay rescans past notes and recognizes the person in future recordings. If you name a speaker with a name that already exists, the two profiles merge as the same person, so duplicates are easy to clean up.
Recognized in later recordings
In later recordings, recognized voices show their profile names instead of generic labels. Before relying on the names across a long note, review the first few passages. Short answers, interruptions, and overlapping speech are harder to assign, and HoldSay leaves those as generic labels rather than guessing.
Named speakers use real photos and remain easy to distinguish in later review.
- Wrong person? Correct the label in the note; the fix updates the profile everywhere.
- Two people sound alike? Keep a separate profile for each and correct the occasional mix-up in the note.
- Manage profiles in Settings under People: rename, change photos, play a sample, or forget someone entirely.
Everything stays on your computer: voiceprints and profiles live in the local HoldSay library and are never uploaded.
Common questions
Does HoldSay recognize my own voice?
Yes. HoldSay saves your voiceprint locally during the first-run test dictation, or backfills it from your recent dictations, and labels you Me in recordings. Recognition follows the same rules for every voice; if you are ever missed, name yourself once and HoldSay remembers.
Can HoldSay remember a speaker's name?
Yes. Name a speaker once in any finished recording and HoldSay recognizes that voice in past and future notes on the same computer.
Are HoldSay speaker profiles stored in the cloud?
No. Voiceprints and speaker profiles remain in the local library on your computer and are never uploaded.
If a label on this page does not match the app, email support@holdsay.com. Include Mac or Windows and the HoldSay version shown in Settings.