Transcribe audio
Turn recordings into editable text quickly so you can find and fix parts faster.
Everything you need
Record, transcribe, edit, and export audio and video — without leaving the app.
Core workflow
Edit your words,
or your waveform
Transcribe any recording and edit the text like a document. Words you delete are automatically removed from the audio — no timeline scrubbing needed.
Detect filler words
Kill the "uhm" and "you know" in seconds
The app detects filler words and speech disfluencies in your transcript and marks them for removal — all running on your device, with no data sent anywhere.
Smart silence detection
Dead air and gaps,
gone in one tap
Automatically detect silent gaps and moments where no one's speaking. Review what's flagged and clear it all at once — no manual scrubbing.
Voice cloning
Fix a word without
re-recording
Patch any word or phrase using on-device voice cloning. Type the correction and the app generates the audio in your voice — seamlessly spliced in.
Frictionless edits
Edit fast with ripple edits
Every cut ripples through the timeline automatically. The transcript, waveform, and video track stay perfectly aligned after every edit.
Built-in recording
Record your screen,
then edit it
Capture your screen with smart auto-zoom, then jump straight into transcript-based editing — no third-party recorder needed.
Features
Everything in the app right now
Edit audio like text
Simply edit text to generate audio. No re-recording required.
On-device AI models
Powered by AI models that run directly on your device for privacy and speed.
Detect filler words
Detect filler words like "uhm", and other speech disfluencies like "you know".
Silence & dead air detection
Automatically spot silent gaps and moments where no one's speaking — clean up pacing in minutes.
Speech-gap detection
Find segments where no one's talking — background audio, pauses, ambient noise — and clear them in one go.
macOS support
Built for Apple Silicon on macOS 15.2 and above.
Export audio and video
Deliver polished outputs once edits are done, without extra tools.
Screen recording
Capture your screen with smart auto-zoom, then jump straight into transcript-based editing.
Upcoming features
Coming soon
Speech enhancement
Improve speech clarity and remove noise from audio.
Coming soon
More AI models
Expanded model support is on the way, including Pocket TTS and Nvidia Parakeet.
Coming soon
MCP
Connect to AI agents to automate editing your media
Coming soon
Microsoft Windows support
Use the same workflow beyond macOS when the Windows app is available.
Coming soon
YouTube chapter generation
Auto-generate clean chapter markers from your transcript for faster publishing.
Coming soon
Captions
Create and export timed captions from edited transcripts in a few clicks.
Coming soon
Export FCP XML
Generate Final Cut Pro XML exports for timeline handoff.
Coming soon
Backgrounds
Change or remove backgrounds
Get a license
One-time purchase. Yours forever.
Unlock exports and full editing features with a one-time license that supports 2 devices on macOS (Apple Silicon).
FAQ
Behind the build
How I built this
A quick breakdown of what it took to build ZxClip.
I built ZxClip for myself, to edit videos fast. I've had fun building this. I hope you love using it. Below are some quick info about how this app was built.
- Tauri with Rust & Swift, alongwith Stimulus.js (yes!) and Basecoat for frontend
- Ported 4 models to Swift using Apple MLX (for MacOS). Whisper, Chatterbox TTS, Wav2Vec2 and Gemma3. This implementation of Whisper comes with some customizations curated from various other implementations.
- Ported 7 models to Rust with Burn framework for Windows version of the app. Whisper, Gemma3, Bert, Wav2Vec2, Vocos, F5 TTS and soundchoice g2p.
- Fine-tuned BERT for a new disfluency classification model. This model is used in the app to detect filler sounds and words. Model & dataset will be on HuggingFace soon
- Fully built with AI, and a few months of shepherding coding agents.