Launching today
Find the exact moment inside any recording, locally on your machine, with no uploads and no cloud. Find what someone said without remembering their exact words. Ugle searches spoken dialogue by meaning across your audio and video archive, returning timestamped passages you can play and verify. Transcription, indexing, and search run on your machine, with no media uploads required and no per-search token fees.
Free
Launch Team
Framer AI AgentsDesign and publish professional sites with AI
Promoted
Hey everyone, I'm Tejas and I built Ugle.
It started with a dumb problem. I knew someone had said a specific thing in one of my recordings, and I couldn't find it. I tried scrubbing through files and searching by filename, then gave up. I had hundreds of hours of audio and video and no way to search what was actually said in it.
So Ugle transcribes your whole media folder on your Mac and lets you search it the way you'd remember it. You type "the part where we argued about pricing" and it takes you to that moment, even if nobody used those words. Click a result and it plays right there.
A few things I cared about:
- It all runs on your machine. No uploads, no cloud, no account to see your files. Your recordings never leave your laptop.
- It works in a lot of languages. Transcription covers 99+ of them. If you're in India, there's an optional Indic pack that handles Hindi, Tamil, Bengali and six more, including the usual mid-sentence switching into English words.
- It's built for people who edit. Collect the moments you find into a clip list, then export it straight to Premiere, Final Cut or DaVinci Resolve. You can also export transcripts as SRT/VTT.
Being honest about where it is today: it's Mac only (Apple Silicon), and the first indexing run takes a while on a big library because it's all happening locally. Multiple library folders and Windows are next.
There's a 15-day free trial. I'm the only person building this, so I'll be here all day reading every comment. Tell me what breaks, what's confusing, or what you'd use it for. Bug reports are very welcome
Most client projects have strict NDAs, so uploading raw footage to cloud transcription services like Descript or Otter has always been an uphill battle with legal. A local-first tool that searches by concept is exactly what our post-team needed.