STINGER VOICE PRINTS
She knows a voice she has heard before.
Once a voice is enrolled, Annabelle recognises that returning person by the sound of it. When she is not sure, she says nothing rather than guess.
What it is
Recognizing a person by the sound of their voice.
Once a voice is enrolled, Annabelle recognises that returning person from the sound of it when they speak.
What it does for you
Knows a known voice
When a person whose voice has been enrolled speaks again, she can recognise that returning voice by its sound. Un-tuned matches fail to “unknown”, not to a wrong name.
Less asking who is calling
A recognized voice means she does not have to start from scratch each time that person speaks.
A hint, never a password
A recognized voice is treated as a clue, not proof. It opens no door and grants no access, and when she is not sure she says nothing rather than guessing a name.
Works with the rest of Annabelle
Knowing who is speaking helps her put the right context on what is said.
No separate screen
It runs while you talk to Annabelle, and your voice stays on your device.
Your device works out a small signature from the microphone it already has open while you speak to Annabelle, and only that signature is sent. The recording itself never leaves the device and is never stored. There is no setup screen to fuss with: a voice is enrolled from the speaking you already do, and recognised on later calls.
How it works
How it compares
Voice Prints against Otter and Granola.
Otter and Granola both label speakers within one meeting. Neither vendor's own page states recognizing the same named person across different meetings and calls, or says where a voice signature is computed and stored.
| Feature | Voice Prints | Otter | Granola |
|---|---|---|---|
| Recognizes a specific known person's voice across DIFFERENT meetings and calls, not just within one | Yes | Not stated on its page | Not stated on its page |
| Where the voice signature is computed | On your own device; only the signature is sent, never the recording | Not stated on its page | Not stated on its page |
| What happens when a voice does not match anyone known | Treated as unknown, never a guessed name | Not stated on its page | Not stated on its page |
| Extra per-seat subscription | Included with every Command Center tier | $8.33-$30/user/mo depending on tier and billing | $0-$35/user/mo depending on tier |
Competitor figures read from each vendor's own page, 2026-09-21. Prices change; re-verify before treating as evergreen.
Included
Included with every Command Center tier.
Get started