Skip to content

Using Heirvo

Build a Searchable Family Video Collection on Windows

If you remember what someone said, a speech transcript can help you find the recording again. Heirvo Archive uses a local Whisper model to recognise spoken English and makes those recognised words searchable. It does not identify people or events from the pictures alone.

Disc sleeves organised in a wooden file holder beside a portable hard drive.
Illustrative image · Heirvo guides

Start with a playable recording

Recover readable footage from your own disc, or use Archive to import a supported media file. Keep the original. Play the recording and confirm that the speech you want to find is audible.

Archive's local vault, importing and speech search are part of the next release's verification. Check the current download page before buying. A Recover license unlocks saving recovered disc files; Archive adds the collection and import workflow.

Find something that was actually said

  1. Open the recording in Heirvo and allow media preparation and transcription to finish. Long recordings can take time.
  2. Use Search every spoken word… and type a word or phrase that was spoken in the recording.
  3. Open a matching result and let the viewer play through the associated passage.
  4. If a name is missing or was transcribed incorrectly, try another word spoken nearby. No match means no recognised matching transcript text, not proof that the event is absent.

What the result does and does not prove

Background sound, overlapping speakers, accents and damaged audio can cause recognition errors. We do not claim a measured word-accuracy percentage for your footage. Search uses recognised text, not automatic face recognition or guaranteed fuzzy matching of misspelled names.

Playback starts at the transcript's associated time. Timestamps can be approximate, especially around long silences. A result that starts early should be checked by listening. No frame-accurate or exact-second guarantee is made.

Media processing stays local

The speech engine, transcript and search index run on your Windows PC. The app does not upload your recordings for transcription. Internet is required for activation, occasional license checks and any missing component downloads. With components installed and a valid cached license, local processing can work offline.

A few more answers

Common questions

Can I search for fishing if nobody says that word?

Spoken-word search does not identify visual scenes. It can find fishing if that word was recognised in the audio transcript. A silent scene needs to be located another way.

Does the app upload my home video?

The recovery and speech processing described here run locally. License checks and component downloads use the internet; they do not upload the recording for processing.