Drop in a recording. The editing is done.
Cutting, captions and multicam switching in one pass. Cuditing cuts what you don’t need for each speaker, captions the edit with names checked by listening again, and cuts to whoever is talking. One click undoes anything.
Free to cut and preview. Sign up and export up to 2 hours a month for free.
The whole edit in one pass
Not just a silence cutter.
Automatic cuts
Silences, fillers, false starts and pre-show chatter, per speaker. Marked in red over the full recording so you can check each one.
Captions
Captions timed to the edit, styled your way and burned in. SRT and VTT too.
Listening check
Two AI models listen again; a misheard name or word is fixed only when both agree.
Multicam switching
Cameras lined up by sound, and the picture cuts to whoever is talking.
Separate audio
Upload the mic audio and the camera video apart; the offset is found and fixed automatically.
Edit the script
Click a word to cut it, click again to bring it back. Fix the text and the captions follow.
How it works
You set the rules. Cuditing makes the cuts.
- 01
Bring your recording
Drop in video or audio. Add more clips, a voice-over or music any time. Long recordings are fine.
- 02
Set the rules
Choose how many people are talking, how long a pause can last, and what to clean up. Each kind of recording starts with its own defaults.
- 03
Review, then export
Every cut is marked on the timeline and in the transcript. Anything uncertain is left in and waits for your decision.
Silence threshold
Drag the line. Watch the silence go.
Every pause longer than your threshold is shortened to a natural breath. Try it on this waveform.
Finishing by hand
Cut automatically. Finish by hand.
Delete the words, cut the video
Drag across words in the transcript and press Delete to cut them. Press R to bring them back.
Margins on both sides
Set how much air to keep after someone stops talking and before they start again, so cuts never feel abrupt.
Fast-forward silence
Keep silences you need but play them at 1.5×, 2× or 4×. Useful for screen recordings and anything people need to watch.
What it takes out
Only what slows you down.
Dead air
Long pauses become a short, natural breath. Not a jump cut.
Filler words
Um, uh, 음, 어. Removed when they stand alone, left in when they are part of a word.
False starts
When someone starts a sentence twice, the first take goes and the clean one stays.
Pre-show chatter
Mic checks and “are we live?” before the real start are trimmed.
And what it keeps.
Short reactions like “right” or “맞아요”, and laughter. A conversation should still sound like one.
Multiple speakers
Tell it how many people are in the room.
Set the speaker count up front and a four-person panel stays four people. Reactions stay with whoever made them, and no pause is cut while someone else is still talking.
Multicam
The camera follows whoever is talking.
Upload the angles of one recording: they are lined up by sound, and each camera is matched to the person it hears loudest. Short reactions don’t switch the picture; quick back-and-forth goes to the wide shot.
- Synced by soundNo clap or slate needed: lined up to a hundredth of a second from the recorded sound.
- Separate mic tooThe mic audio leads; only the picture comes from the camera.
- Your pick anywhereClick the camera row on the timeline or press a number key.
···
Captions
Cut words leave. The rest become captions.
Hosttoday we’re talking about why small teams move fast.
Captions
The cut is done. So are the captions.
Words you cut leave the captions too, and what stays is retimed to the edit. Misheard names and words are fixed by listening again, and the captions are burned in with the weight, size, position, colours and background you pick. All on the free plan.
Frame-accurate to the edit
Caption times come from the same timeline as the rendered file.
No “um” in the captions
Fillers and false starts leave the captions along with the cut.
Looks the way you set itFree
Pick the weight, size, position, text colour, background colour and opacity with a live preview, and export it burned in.
Hosttoday we’re talking about why small teams move fast.
- Sponsor readsIntroducing a sponsor or a product
- Internal talkCompany schedules or talk about colleagues
- Add rule
Before we get going, one quick thing.
This episode is brought to you by Keychron keyboards.
Oh, and Friday’s meeting moved to three.
Right, let’s start with the first question.
Your own cut rules
Describe what to cut. In a sentence.
Write down what you never want in the edit. Cuditing reads each sentence in context and finds the matches: clear ones are cut, borderline ones wait for your review.
Up to 3 on Pro, 10 on Ultra. Saved with your style, so the inbox applies them too.
Who it’s for
Anyone who talks for a living.
Podcasters
Two hosts, a guest, an hour of tape. Get to the version you would actually publish without scrubbing through every pause.
Silence over 1.0s · fillers · restarts
Game streamers
Turn a long broadcast into an edit you can pull highlights from. Gaps are measured by your voice, not by volume, so game audio and music don’t count as talking.
Voice gaps over 3.0s · keeps laughter and reactions
Interviewers
Keep the answers, lose the scheduling talk, and get tangents flagged for you to decide. Built for multi-person sit-downs and remote calls.
Silence over 1.2s · suggests tangents
Educators
Lectures and tutorials without the long pauses while you think or type. Students get the explanation, not the wait.
Quiet over 0.8s · stricter fillers
Exports
Finish in the editor you already use.
Final Cut Pro
FCPXML with every cut as an edit point.
Premiere Pro
XML timeline, ready to relink and refine.
Subtitles
SRT, WebVTT and video with captions burned in, retimed to the edit. Speaker names and your own font, colours and background on every plan.
Finished file
MP4 or M4A with your voice-over and music mixed in.
Cuditing Cloud
Record. Drop it in. The cut is waiting.
Your recordings live next to the editor, not on your desktop. Every plan comes with cloud storage; the inbox and review links start on Pro.
Drop files, get cuts
Every file you drop becomes a project, cut with your saved style.
Your rules, saved
Speaker count, language, silence threshold and what to keep. Set them once for each show and reuse them.
Send a link, not a file
Co-hosts and editors play the edit in the browser, accept or restore cuts and leave notes at the exact second.
Up to 2 TB
Originals and renders on Cloudflare R2. Downloads come straight from storage, at full speed.
Pricing
Editing is free. Sign up to export.
Sign up with just an email for 3 hours of auto cut and 2 hours of exports every month, free. Pro and Ultra open soon.
Free
Try it on a real recording, then export from your monthly hours.
- 2 hours of exports a month
- Free with an email sign-up
- 3 hours of auto-cut a month
- 5 GB cloud storage
- Projects kept 7 days after last edit
- Edit captions, speaker names in SRT
- Burned-in captions with speaker names, colours, size and position
- Multicam: up to 2 cameras, switched by speaker
- 1 saved styles
Pro
For a weekly show or a regular stream schedule.
$228.00 billed yearly
- 12 hours of exports a month
- Then $0.04 per minute
- 36 hours of auto-cut a month
- 100 GB cloud storage
- Projects kept while subscribed
- Inbox: drop files, get cuts
- Review links for co-hosts and editors
- Edit captions, speaker names in SRT
- Burned-in captions with speaker names, colours, size and position
- Your own cut rules: up to 3
- Word list: cut exact words
- Multicam: up to 3 cameras, switched by speaker
- 10 saved styles
Ultra
For long streams, daily uploads and small teams.
$756.00 billed yearly
- 50 hours of exports a month
- Then $0.03 per minute
- 150 hours of auto-cut a month
- 2 TB cloud storage
- Projects kept while subscribed
- Inbox: drop files, get cuts
- Review links for co-hosts and editors
- Edit captions, speaker names in SRT
- Burned-in captions with speaker names, colours, size and position
- Your own cut rules: up to 10
- Word list: cut or bleep
- Multicam: up to 6 cameras, switched by speaker
- 50 saved styles
- Priority processing
- 3 seats
Questions
Can I keep silences instead of cutting them?
Set Silences to “Fast-forward” in the rules and they play at 1.5×, 2× or 4×. The preview, the finished file and the captions all follow. Final Cut Pro and Premiere Pro exports carry the start of each such range at normal speed.
Can I change what counts as silence?
By default the level is set from the recording’s background noise. Turn off automatic detection to pick anything from -65 dB to -20 dB.
Does it change my original file?
No. Cuditing keeps a list of cuts and leaves the source untouched. Restore any cut at any time, even after exporting.
Which languages does it understand?
Korean, English, Japanese and Chinese, including conversations that mix Korean and English. Choose the language up front or let Cuditing detect it.
Can I cut specific kinds of talk?
Yes. On Pro and Ultra, describe what to cut in a sentence, like “sponsor reads” or “internal talk”, and Cuditing finds it and cuts or suggests it. Exact words go in a word list.
Do captions show who is speaking?
Yes. WebVTT always carries speaker names, and you can add them to SRT and to burned-in captions. Fixing caption text, choosing the font, colours and background, and exporting video with captions burned in work on every plan, free included.
How does it handle several people talking?
You choose the number of speakers before the cut. Cuditing separates their voices, never cuts while anyone is still speaking, and keeps short reactions from the other side.
What if it cuts something I wanted?
Cuts it is unsure about are only suggested, not applied. Every applied cut is listed with its reason and can be restored with one click.
Can I use it for game streams?
Yes. The streamer preset measures gaps in your voice, not in volume, so game audio doesn’t hide the stretches where nobody is talking.