Does Wave remember voices between different recordings?

Yes. That's what makes Voice ID different from standard speaker separation, which starts from scratch each time. Voice IDs live on your Wave account and sync everywhere, so a voice saved on your phone is recognized in recordings made on the web or the Wave desktop app too.

Does voice identification work for phone calls and imported files?

Yes. Wave's Voice ID works on any session with separated speakers, including meetings, phone calls recorded through Wave, in-person recordings from your phone or watch, and imported audio files.

What is People in Wave?

People is Wave AI's list of everyone you've saved a Voice ID for. Each person has a page with every recording they're in, how long they spoke and how Wave identified them, and you can rename, merge duplicates, mark yourself with This is me, or forget a voice. It's on iPhone, Android, the web and the Wave desktop app.

How accurate is Wave's speaker identification?

Wave needs at least 10 seconds of a person's speech to match them, and it leaves a speaker unlabeled rather than guess on a weak match. Clear audio and less crosstalk help. If Wave names the wrong person, rename the speaker or choose Not them on that recording, and Wave stops learning their voice from it.

Is my voice data private?

Yes. A Wave Voice ID holds a numeric voice fingerprint and a short audio clip, used only to identify speakers in your recordings. You can delete any Voice ID at any time with Forget voice or turn automatic voice matching off, and Wave does not use your recordings to train AI models.

Voice ID: Wave Remembers Voices From One Call to the Next

Q: How does Wave identify who is speaking?

Wave AI separates speakers automatically in every recording. When you name a speaker and save them as a Voice ID, Wave stores a voice fingerprint and matches it in later recordings, so that person is labeled by name automatically instead of as Speaker 1.

Last updated: November 3, 2026

Wave AI is a voice identification app for iPhone, Android, the web, Mac and Windows: it separates every speaker in a recording, and once you save someone's voice as a Voice ID, it names them automatically in every future meeting, call and voice memo. People then lists everyone Wave knows by voice, with every recording they're in.

How Voice ID and People work

How does Wave identify who is speaking?

Wave splits every recording into separate speakers. Name a speaker and save them as a Voice ID, and Wave stores a voice fingerprint. In later recordings it matches voices against your saved Voice IDs while the recording processes, so names are in place before you open it.

Record or import a conversation. Wave labels Speaker 1, Speaker 2 and so on.
Open Speakers, play each voice, and type a name, or let Wave suggest names from the conversation.
Save them as Voice IDs, and mark your own voice with This is me.
From then on, Wave labels those people automatically in new recordings.

Wave help: Voice ID

People: everyone Wave knows by voice

People is the directory of your saved voices, on iPhone, Android, the web and the Wave desktop app. Each person has a page with every recording they're in, how long they spoke, and how Wave identified them.

Rename, merge duplicates, or Forget voice to delete a Voice ID for good.
Not saved yet: speakers from the last 30 days who aren't linked to anyone, ready to save or link.
Same person? Wave flags two saved names that sound alike and lets you listen and decide.
Not them: fix a wrong match on one recording without touching the rest.

Wave help: People

How accurate is it?

Matching is deliberately conservative. A speaker needs at least 10 seconds of audio in a recording, and Wave leaves someone unlabeled rather than commit to a weak match. Profiles improve as Wave recognizes people with high confidence, and you can switch that learning off per person.

The problems you know too well

'Speaker 1' Isn't a Name

Generic speaker labels make transcripts hard to use; you're re-identifying the same colleagues in every recording. Voice ID replaces anonymous labels with real names.

Every Tool Forgets Between Calls

Typical speaker separation starts from scratch each recording. Wave remembers: once you save a voice, that person is recognized in your next meeting, next week's call, and every conversation after.

Attribution Matters

Meeting minutes, interview quotes, and commitments all depend on who said what. Persistent voice identification makes attribution automatic instead of an editing chore.

Features

Everything you need to capture every conversation

Voice ID Across Recordings

Save a voice once and Wave recognizes that person automatically in future recordings, on every device, because Voice IDs live on your account.

Automatic Speaker Separation

Even for brand-new voices, Wave separates and labels speakers in every transcript, and can suggest names from the conversation itself.

Works Everywhere You Record

Voice ID works on any session with separated speakers: meetings, phone calls, in-person recordings from your phone or watch, and imported audio files.

A Page for Every Person

People shows each saved person with every recording they're in, their talk time, and their key lines, so you can find a conversation by who was in it.

Ask Your AI Tools About People

Connect Wave to Claude or ChatGPT and ask "What meetings was Sarah in this month?" The assistant can list the people you've saved and find the recordings each one spoke in.

Private by Design

A Voice ID is a numeric voice fingerprint plus a short clip, used only to identify speakers in your recordings. Delete any of them at any time, or turn automatic voice matching off.

Perfect for

Recurring team meetings with the same voicesClient calls where attribution mattersInterviews and panel discussionsBoard meetings and formal minutesSales teams reviewing who committed to whatResearchers tracking participants across sessions

FAQs

Common questions

The Wave Guide to Meeting Recording

Four ways to record any meeting with Wave — meeting bots, desktop app, phone calls, and mobile. Pick your method and never miss a detail.

All The Ways to Wave: Every Device, Every Scenario, One App

The complete guide to recording with Wave — iPhone, Android, Apple Watch, Wear OS, Desktop, Chrome Extension, Meeting Bots, Phone Calls, Web App, and Import. Every way to capture, transcribe, and summarize your conversations.

Introducing Voice ID: Wave Learns Who’s Talking

Voice ID creates a mathematical fingerprint of each speaker’s voice, then automatically identifies them in every future recording — making transcripts and summaries more useful without any manual labeling.