Find a specific person in your indexed video library. Preview a detected voice, check its source clip and assign it to a name. Search for matching footage, including moments where the person speaks off camera. Face and voice recognition run on your device.
Optional and consent-gated. Available on every plan including Free.
Works with
your workflow.
Exports to your NLE. Reads cinema RAW natively.
Assign matching faces and voices to the same person. That name becomes a search filter across your indexed footage, including new clips analysed with recognition enabled. Review matches as you work, especially when recording conditions change.
Sarah
3 faces · 1 voice
James
5 faces · 2 voices
Right-click or drag onto a person to assign
Voices 2
Interactive illustration · no audio0:42–0:47
0:00 / 0:05
1:18–1:23
0:00 / 0:05
A voice you can check before you name it.
Drag a waveform or try Assign with the example person, Sarah Chen.
Face and voice templates are created and stored on your computer
Face and voice recognition are optional and off until you turn them on. Models download to your machine and run locally. The templates they create are encrypted at rest, and FrameQuery's servers never receive them or the names you assign. Frames and audio uploaded for indexing can show faces and contain voices; they are not used to recognise anyone. How face and voice data is handled.
Face and voice recognition run on your computer. Templates and the names you assign are never sent to our servers.
Face and voice templates and face thumbnails are encrypted with AES-256-GCM. The key is held in your operating system’s keychain or credential store.
Off until you turn it on. Face and voice are separate choices, and you tick a consent box first.
Templates and names are never included in backups, shared indexes, exports or reviews. “Delete all biometric data” removes all of it, and unused data is deleted after one year.
A person can have several linked faces and voices. Use @"Sarah Chen" to find their matching footage, then add a quoted phrase or a visual detail to narrow it. The current project and search filters still apply. Merge duplicate people when two names refer to the same person.
Search by the person you named
Interactive example@"Sarah Chen"interview-take-04.mp4
0:42“The river was here long before the road.”
Face and voice match
location-interview-02.mp4
1:18“We came back the following spring to film it.”
Voice match · speaking off camera
The Voices section in Manager → People shows source filenames, spoken intervals and audio waveforms. Scrub to a useful moment, listen, and open the source clip if you need more context. Assign the voice to an existing person or create a person directly from the card.
Play, pause and scrub the waveform. The filename opens the same interval in its source clip.
Give matching voices and faces one name. Names stay useful even when the person speaks off camera.
Use Search videos from the card, or an @name query. An unnamed voice can be searched directly too.
Diarization separates speakers within a transcript. Recognition matches voices across recordings so you can associate them with a person. For the spoken words themselves, use transcript search.
Previews need an accessible source clip or usable local proxy; a missing preview is not evidence that no matching voice exists. Follow the voice preview and assignment steps.
Available on every plan including Free, because it runs on your device and uses no indexing hours. It works on indexed video, from your own indexing or a shared index. Enable when you’re ready and FrameQuery backfills any previously indexed videos.
Privacy, consent, searching by person and deleting biometric data.
Search every word spoken across your entire video library.
Learn moreDescribe what you're looking for in plain language.
Learn moreShare indexes and back up to local, Drive, or NAS.
Learn moreFind every near-duplicate of a frame or clip across your own library.
Learn more
Preview a voice, assign a name and search matching footage by person.
Runs on your device. Available on every plan including Free.
