Become an affiliate. Earn 20% recurring commission on every referralEarn 20% recurring commissionEarn 20% recurringApply Today
New · Premiere Pro Extension

Editor

Your assistant editor inside Premiere Pro.

Editor transcribes your sequence on your own machine, then cleans a talking-head cut in one pass: silences, breaths, filler words, profanity, and repeated takes. You read every change in the script before anything is built.

Pr
Start Free Trial
✦ 7-day free trial

Editor comes with Filmit Studio, along with our whole suite of plugins, effects, transitions, overlays, and courses, plus everything new we ship. It's all included!

Included with Filmit Studio
v1.0
Setup Cuts Review

Watch the Demo

See Editor in Action

A full walkthrough, from a raw 37-minute recording to a clean cut, including the calls it gets wrong and how the review catches them.

Editor Demo
Get Started

How to get Editor

Three steps and you are up and running, with Editor and every other Filmit tool, right inside your editor today.

1

Create your account

Sign up free at filmit.io. No card needed to look around.

2

Install Filmit Studio

Download the free desktop app for Mac or Windows. It manages every tool for you.

3

Start your free trial

Start your 7-day trial and install Editor plus every other tool, straight into your editor.

The work before the work.
Every talking-head edit starts with the same unpaid hour, and none of it is the part you got into this for.

Scrubbing for dead air

Gaps, breaths, and half-second pauses, one at a time, up and down a sequence you have already watched twice.

One pass marks all of it.
👀

Cutting blind

Automatic tools hand you a finished sequence and hope you agree. If a cut was wrong, you find out on the timeline.

You read the script first.
🔨

Jobs that are not editing

Rolling J and L cuts by hand, switching angles, evening out the mix, exporting the same cut at every size a client asks for.

One click each.
Transcribe. Mark. Review.
Three stages between a raw talking-head sequence and a clean one, and you have the last word at every step.
Stage 1 · Transcribe

Transcribed on your machine.

Editor exports the audio from your sequence and transcribes it locally with the same Whisper engine we share across our tools, using your GPU when there is one available. Nothing is uploaded, and every word carries its own timing.

  • Runs on your own machine, no upload
  • Uses your GPU when one is available
  • Word-level timing is what the cuts are built from
  • One engine, shared with our other transcription tools
  • Pick the audio channels that carry the voice, so music stays out
  • A recovery pass re-listens to stretches with sound but no words
Local Transcription
0:06
0:14
0:23
Stage 2 · Mark

One pass, every rule you ticked.

Tick what you want gone and Editor marks all of it in a single pass: silences, breaths, filler words, profanity, and takes that were started again. Profanity is yours to handle, with a bleep, a mute, or a cut.

  • Silences, breaths, filler words, repeated takes
  • Profanity bleeped, muted, or cut
  • Each kind of mark has its own colour
  • Word lists per language live in Settings
  • Silences: you set how long, and how quiet, a gap has to be
  • Relaxed keeps laughs and reactions; Aggressive also cuts the ums Whisper skipped
  • Not in script catches restarts and stutters the transcript left out
  • Bleep with a tone or your own sounds, one or several at random
Marked Script Marked
Stage 3 · Review

Read it before it is built.

Review mode shows you every changed passage with the words around it, so you judge a cut in context instead of guessing from a waveform. Click a mark to keep it. Select a span of words to cut it yourself. Bring words back inside a cut you already have. Then build, and the backup lands in an Editor/Backups bin.

  • Changed passages shown with their context
  • Keep any mark with a click
  • Cut or restore any span of words by hand
  • Backups filed in an Editor/Backups bin
  • Full script, or Review mode with only the passages that change
  • Jump the playhead to any line from its timecode
  • Cross out removed words, or keep them readable while you listen
  • Drag lines to reorder them, then build a clean copy under Editor/Cuts
  • Export the marked script and a timecoded cut list for a client
Built Sequence Built · backup kept
V1
A1
A second read, if you want one.
The rules do the mechanical work. An AI read is there when you want a second opinion on which takes to keep, and it is entirely optional.
💬
Free · no key

Export, paste, import

Editor writes the script out ready to paste into Claude or ChatGPT. Paste the answer back into the panel and Editor imports the plan, applying the vetoes and questions to the marks you already have.

One click · your own key

Read it in the panel

Add your Anthropic or OpenAI key once in Settings and Editor runs the read for you. The AI can veto a cut the rules made, or flag a passage it is not sure about, and you still decide in Review.

Either way, the transcription stays on your machine. The only thing an AI ever sees is the script text you choose to send. Never your video, never your audio.

The rest of the assistant.
The jobs that come after the cut is clean, each one a tab in the same panel.
Switcher

Camera angles, cut in place.

Stack your angles on V1, V2, and V3 and Switcher cuts between them for you. Each angle holds for a random stretch inside the range you set, so a two or three camera interview becomes something a client can actually watch. Your audio is never touched.

  • Hold each angle between a minimum and maximum number of seconds
  • Start on a chosen angle, or let it pick at random
  • Switch the whole sequence, or only the clips you select
  • Disable the off angles, or remove them outright
  • A backup copy is kept before anything changes
Switcher
V3
V2
V1
A1
Audio untouched · backup kept
J and L cuts

Every edit point, rolled at once.

One J or L cut is easy by hand. A whole interview of them is an afternoon. Select your clips and Editor rolls the audio cut at every edit point between them: J to bring the next shot’s sound in early, L to let the last shot’s sound carry over, or alternating so a run of cuts does not feel mechanical.

  • J, L, or alternating down the timeline
  • Offset in seconds or frames
  • Optional Constant Power crossfade at the moved audio cut
  • Works on butted clips and on checkerboarded audio
J and L cuts
V1
A1
L cut · 0.5 sConstant Power
Audio

A mix that lands at the right level.

Render the sequence mix, loudness-normalize it to a standard target, and drop the result onto a fresh audio track, filed next to your project. Noise removal is there too, with a strength you choose, and it is the more experimental of the two.

  • Targets for YouTube and social, podcasts, and EBU R128 broadcast
  • Optional low-rumble cut, a high-pass at 70 Hz
  • Noise removal with a strength you choose
  • Processed audio lands on its own track, your originals stay
Loudness
-14 LUFS
YouTube · Podcast · EBU R128High-pass 70 Hz
Sequencer

Every delivery size, in one pass.

Pick clips or finished sequences and tick the sizes you need. Sequencer builds one sequence per source for each: vertical 9:16s from 4K footage, 16:9 4K, 2.39:1, 8K, or a frame size of your own, scaled to fill or fit. Each run is filed in its own folder under Editor/Sequencer.

  • From 9:16 to 8K, plus frame sizes of your own
  • Scale to fill the frame, or fit inside it
  • Frame rate matched to the source, or set your own
  • Pixel aspect follows the footage
Sequencer
9:16
1:1
4:5
16:9
2.39:1
Fill or fitEditor/Sequencer
🗑

Backups by default

Editor builds next to your work, never over it. Backups land in an Editor/Backups bin, so the sequence you started with is always still there.

📋

Quiet by design

Every tab shows a single status line while it works. The full activity log lives in Settings, along with the word lists and the colour for each kind of mark.

🎨

Your colours, your sounds

Pick a colour for every kind of mark so the script reads the way you think, and load your own bleep sounds for the profanity pass.

From raw sequence to clean cut in four steps.
No paper edit, no logging, no API key needed to start.
1

Pick your sequence

Open Editor from the Filmit Studio panel in Premiere Pro and point it at the talking-head sequence you want to clean.

2

Transcribe

One click. The transcription runs on your own machine and every word gets its own timing.

3

Tick your rules

Silences, breaths, filler words, profanity, repeated takes. Add an AI read if you want a second opinion.

4

Review, then build

Read the changes in the script, keep or cut by hand, and build. Your original is backed up in an Editor/Backups bin.

Common questions
Everything you need to know about Editor.

Adobe Premiere Pro, as a native panel on Mac and Windows. It reads the sequence straight from your project and builds the cleaned version back into it. Editor is Premiere Pro only for now.

No. Transcription runs locally with our shared Whisper engine, on your own machine, using your GPU when one is available. Nothing is uploaded. If you use the optional AI read, the only thing sent is the script text you choose to send, never your video or audio.

No. The rules do the cleanup on their own. If you want a second opinion, Editor exports the script ready to paste into Claude or ChatGPT and imports the plan you paste back, with no key at all. Adding your own Anthropic or OpenAI key just moves that round trip inside the panel.

Yes, that is the point of Review mode. Every changed passage is shown with the words around it, so you read each cut in context. Click a mark to keep it, select a span of words to cut it yourself, or bring words back inside a cut you already have. Nothing is built until you say so.

It stays exactly as it was. Editor builds next to your work and files backups in an Editor/Backups bin, including before a Switcher pass. You can always go back to what you started with.

Switcher cuts between stacked camera angles in place. J and L cuts roll the audio cut at every edit point, earlier, later, or alternating, with an optional crossfade. Audio loudness-normalizes or denoises the mix onto a track you choose. Sequencer fans any clip or sequence out to every delivery size and format you tick, filed into named folders.

Relaxed only cuts words that are in the transcript, so laughs, reactions, and short sounds like ‘hm’ stay in. Aggressive, the default, also cuts the ums and uhs the transcription skipped, marking them straight from the audio. Pick Relaxed for interviews where the reactions matter.

Yes. Export script saves the transcript with every planned cut marked in your colours, plus a cut list with timecodes, as an .rtf that opens in Word or as plain text. Send it for approval, then build.

Yes. Editor is included with your Filmit Studio subscription at $20/month, along with every other Filmit tool, effect, transition, and course. Try it with a 7-day free trial.

✦ 7-day free trial

Hand the first hour to your assistant editor.

Every tool, every effect, every course. One subscription, 7 days free, then $20/month.

Start Free Trial