Scrubbing blind.
Dragging through a long recording hoping to land on the ten seconds that mattered — and overshooting every time.
Audio
Find the sound, the song, or the words someone said.
Voice memos, recordings, the audio inside your videos — hours of it, and the moment you want is somewhere in the middle. Describe the sound or the sentence, and LatticeSearch takes you straight to it. On your device, always.
Spoken words and sound events alike — timestamped, private, found in ~250 ms.
Heard on your device · jump to the exact second
Sound familiar?
You can’t skim audio. That’s the whole problem.
A photo you can glance at. An hour of recording, you have to sit through.
Which recording was it?
Dozens of near-identical voice memos. The one where they said the thing is in there — somewhere.
The sound in the video.
The bark, the applause, the song in the background — buried in the audio track of a video you’ll never scrub through.
The fix
Search what was said — and what was heard.
LatticeSearch transcribes speech and recognises sounds on your device, so a spoken phrase or a sound event becomes a timestamp you can jump to.
Heard on-device · sound events + speech, timestamped
Spoken words
Find the exact sentence in an hour of talking.
Every recording and video is transcribed privately, so you can search speech by what was actually said — and jump straight to the moment.
- Search meetings, memos, and interviews by phrase
- Land on the timestamp, not the whole file
- Transcription happens on your device
Sound events
Find the bark, the bell, the applause.
Not just speech: LatticeSearch recognises the sounds themselves — a dog, a doorbell, a crowd, a piano — and marks exactly when they happen.
- Search by the sound, not the filename
- Find cues inside video audio too
- All recognised locally, nothing uploaded
Take it apart
Split any track into its parts — and search those.
A mix is a blur. Ask for “that snare” and it drowns under everything else that’s loud. LatticeSearch can separate a recording into drums, bass, vocals and everything else — so each element becomes something you can find, play and export on its own.
- Pull the drums, the bass or the melody out of any track
- Isolate speech from a noisy recording — or strip the voice and keep the room
- Find a track by one element: that bassline, that snare
- Separated on your machine, then cached — instant every time after
Private by design
What you record stays yours.
Voice notes are personal. Nothing is uploaded and there is no account — every transcription and match happens on the machine in front of you.
- 100% on-device
- No cloud, no data collection
- Free to start on the audio you already have
The words you want to hear again
Somewhere is a voice you’d love to hear right now.
A message from someone you miss, the advice you meant to write down, the song you couldn’t name. Ask for it — and there it is.
Find the sound, not the filename.
Private, on-device search for everything you’ve recorded. Free to start — nothing uploaded. Cancel anytime; you keep your files and your index.