Record your screen.Edit the um words.Export the video.
Cutscript records your screen, camera and voice, then turns the take into a script. Delete a word and the video is cut to match. When it reads right, export a file that is ready to post.
The AI runs on the ChatGPT or Claude plan you already pay for, or on models on your own computer. Transcripts need a key, or Whisper on an Apple silicon Mac.
Free account. Free exports have a watermark. Pro removes it from $4.17 a month, paid yearly.
macOS 15 to record, Apple silicon or Intel. Windows 10 or 11, from the Microsoft Store. 0.1 beta.
Click the stickers. That is how you edit.
Step 1 · Record
Press record. Every part of the take stays editable.
Picture, voice, system sound, camera and pointer are saved as separate tracks. After you stop, you can still move the camera, turn down the sound and, on a Mac, restyle the pointer.
Try it. Pick a mode, press start, then click around. This demo records nothing.
1920 × 1080
Not in screen capture
Display 1
Take saved
Every track is on its own, and it opens as a project.
Picture 2Its own file, side by side or full frame.
Picture 3Up to three pictures in one take.
CameraChange its shape, size and place later.
MicrophoneCut it by words. Clean up the voice.
System soundTurn it down without touching your voice.
Pointer and clicks MacRestyle it later. Zooms follow your clicks.
Drawings MacPart of the picture, or kept apart in a window take so you can change them.
Area, Device, drawing and the separate pointer track are Mac only. On Windows the pointer is part of the picture.
Countdown and pauseA countdown of 3, 5 or 10 seconds, or none. Pause and carry on any time.
30 or 60 fpsThe most it will capture. 60 is the default.
Recover a takeAfter a crash or restart, the recorder offers to bring it back.
OfflineRecording needs no internet connection.
Sound from one app MacExperimentalTick the apps you want and leave the rest out.
iPhone and iPad MacExperimentalRecord the screen of a phone on a cable, with its sound.
Ready when you stopThe take opens as a project, with a transcript if you choose, and zooms from your clicks on a Mac.
A clean desktop MacHide your desktop icons while you record.
Record up to three pictures in one take.
Press Add source for another display, window, area or cabled iPhone or iPad. Your camera stays as a bubble on top. The editor opens with a layout called All sources, and Show one source lets you pick which picture fills the frame at any word.
On Windows, added sources are displays and windows.
Draw on the screen while you talk. Mac
Press Draw in the bar during a take. Circle a number, highlight a line, point with an arrow. Screen drawings fade after 3 seconds, or stay until you clear them.
Pen
Highlighter
Arrow
Line
Rectangle
Ellipse
Step counter
Spotlight
Laser pointer
Privacy box
Zoom mark
Whiteboard
A privacy box arrives in the editor as a blur you can change. A zoom mark becomes a zoom.
WithoutDrawn in the take
Step 2 · Edit
Edit the video by editing its script.
Every take becomes a transcript, and so does any video or audio file you open. Select words, press Delete, and the picture and sound are cut to match. Every cut can be restored, and your original file never changes.
A transcript comes from your own AI account, or from Whisper on an Apple silicon Mac. What that costs
Take out the ums in one go.
Remove filler words starts from a list of 17, such as um, uh, you know and basically, and you can change the list. These clean-ups run on your computer, cost nothing and call no AI.
Try it. Click any word to cut it.
0 words00:30.0
Maya Chen
Clean up
Nothing cut yet.
Correct
Fix a misheard word without changing the sound. Correct all fixes every copy.
Ignore
Cut a word but keep it struck through, so you can review it later.
Search
Find a phrase in the script, then delete every match.
Make a plain recording look finished.
A new recording opens with the Studio look: a background, padding, rounded corners and a soft shadow. Pick another, or set each part yourself, for the whole video or one clip. What you see in the preview is what exports.
As capturedStudio
Screen
Look
Also here
Gradients, your own image or a colour
Your Mac’s wallpapers Mac
Padding, corners and shadow
Crop, and 7 aspects from 16:9 to 4:5
Zooms that follow your clicks. Mac
Cutscript adds zooms right after you record, from where you clicked, with Fewer, Balanced or More. On any video, on Mac or Windows, press Z to add one by hand.
Try it. Click anywhere on the frame.
Zoom
Depth
How it moves
Fixed or Follow cursor
Depth up to 5×
A pointer you can change later. Mac
Your pointer is recorded apart from the picture, so you can restyle it after you stop: its look, its size, how it moves and what a click does.
Try it. Move your mouse over the frame, and click.
Cursor
Style
Size
Movement
Click effect
Plus five click sounds: Soft tap, Mouse click, Mechanical key, Pop and Tick.
Nine shapes for your camera.
Pick a shape, set the size and drag it where you want it. Add a border, or chroma key for a green screen, and go full screen or hide it over any words you select.
Try it. Pick a shape, then drag the bubble around.
Webcam
Shape
Size
Remove the room behind you, with no green screen. Pro
Remove background runs on your computer, with no AI account, and nothing is uploaded. Use it on your camera, your main video or any B-roll clip with a person in it, then choose what goes behind them.
Try it. Pick what goes behind her.
Transparent
0:02
Remove background
Behind the person
She stands on whatever is under the video, here a screen recording. A close-up like this covers a lot of it.
Also here
Sharpen edge and Shrink or grow
Standard, or Draft at half the frame rate
Stop it and continue later
It takes time and disk space. Hair and busy backgrounds can leave soft edges or bits of the room, and clear objects can look see-through. These clips are the app’s own exports with no retouching, and they are close to the best case: one person against a plain wall. We did set three sliders by hand: Sharpen edge to 30% and Shrink or grow to -25%, which thin a pale glow round her hair, and Blur’s Strength to 30%. With them where they start, at 0%, 0% and 50%, you will see more of the glow, which shows most on a dark colour, and a wider smear on Blur.
Two, three or four panels, over the words you choose.
Show the screen, the camera, or any video or image in the project, then change the layout at any word.
Layouts
Auto, Rows, Columns, Grid, Large left or Large top
Each panel fills or fits
Grade the footage. Read the scopes.
The Colors panel changes the recording only, never your background, camera bubble or graphics. Or ask the AI: it looks at frames of your video, chooses a grade, then looks again and corrects anything that is off.
Try it. Pick a preset, then drag across the frame.
As recordedTeal & amber
Colors
Presets
Each tile shows your own frame.
Light and color
Exposure, contrast, white balance, vibrance, three colour wheels and a vignette.
.cube looks
Import a LUT from any editor and set how strongly it shows.
4 scopes
Waveform, Parade, Vectorscope and Histogram, in a window you can drag anywhere.
The face is a layer. Click it, colour it, pull it out.
Cutscript finds the faces in your video on your own computer, for free. Click one in the preview to give it its own colours, inside a soft oval that follows it. Or drag it out of the camera corner, and it becomes a bubble that follows the person, with its own Face track on the timeline.
Detaching is for video where the camera is part of the picture, such as a call recording. A take recorded with Cutscript’s own camera already has a bubble.
As recordedFace: Bright
Face
Layer
The face becomes a bubble of its own that follows the face. The screen stays as it is.
Also here
Presets that show the face
Its own light and colour
A Face track you can trim
“Brighten my face” works in plain words
Blur a face once. It stays blurred as it moves.
Blur all faces finds the faces in the video, lists them so you can check each one, and adds a blur that follows it. It runs on your Mac, with no AI account.
Press play, then drag across the video.
As filmedBlur all faces
0:00
Blur
Blur all facesMac
Blur all emails, and phone numbers if you tick the box Mac
By hand: drag a Blur or Pixelate box over anything
Track it forward, backward or both ways, and fix any frame by dragging
Pixel lock for screens, on Mac and Windows. Apple Vision for camera footage, on Mac
If a blur loses what it follows, the export warns you and shows you where
Share an exported video, not a project package: a package still holds the original
Captions word by word, in your style.
Captions come from your script, so a deleted word is never captioned. Choose from 24 styles, or set the font, colour, position and animation yourself. Burn them in at export, or send a subtitle file.
Captions
Style
8 of the 24 styles.
One take, in four languages.
Every recording becomes a script in the language you spoke. Dub this part… then says it in another language over the same picture. The new voice lands on the timeline as an ordinary audio clip, and your original stays as it is.
Try it. Press a language, with your sound on. Click a word to jump to it.
0:00
Video
English
Dub & re-voice
Script
One AI voice speaks all four. It is not her voice, and her lips do not follow it.
This demo is a stock clip with no sound of its own and one AI voice speaking each line. Word timings are from a real transcription of each line. To keep a real speaker’s voice, dub their recording with ElevenLabs Dubbing.
Transcripts
Scribe v2 for 90+ languages, or Whisper on an Apple silicon Mac for 99.
Keep the speaker’s voice
ElevenLabs Dubbing keeps each speaker’s voice, tone and background sound. Copying a voice needs that person’s permission, and the panel asks you to confirm it.
Line by line
Translates your script and speaks each line where it was said, in a voice you pick.
Your voice, a new voice, another language.
Enhance voiceClean up your voice with one of four engines. Built-in clean-up and Apple Voice Isolation (Mac) are free and stay on your computer.
Dub and re-voiceSay a part or the whole video in another language. ElevenLabs Dubbing keeps each speaker’s voice.
VoiceoverType it and hear it, with ElevenLabs, fal.ai, OpenRouter or OpenAI voices, or a voice on your Mac.
Music and sound effectsGenerate them from a sentence with your own account, and see the price first.
Video from a linkPaste a link from YouTube, TikTok, Vimeo and more, for videos you have the right to use.
Text on screenAdd text, arrows, boxes and images. Or let a key line fill the frame, word by word, as you say it.
A real timeline when you want one.
The script and the timeline are the same edit, so use whichever is faster. Add video, image, audio, music and sound effect tracks, with Razor, Slip, snapping, markers and speed changes.
Transitions
10 for video, from Cross dissolve to Whip pan, and 2 for audio.
Compositions
Several edits of the same footage, such as a long cut and a short.
Notes
Notes between lines of the script that never reach the video.
Audio only
A podcast with no picture gets a canvas, captions and graphics, and exports as video.
Ask for a graphic in your own words.
Open Graphics and say what you want. The design director plans it first, so you see a storyboard before anything is built. Press Build, and it makes the graphic, checks it and improves it. Then add it to your video.
Try it. Pick an idea or type your own, then press Send.
Preview
Graphics
Video
GraphicsCreate
Example replies with real graphics from the design director. Nothing runs on this page. In the app, the AI writes a new design from your words, using a ChatGPT or Claude plan or an OpenRouter key.
Bring your files
Attach images, SVG, CSV, Excel, JSON, PDF, Word or text.
Your brand
Pick a brand kit, and it uses your colours, fonts and logos.
Free versions
Usually Quiet, Balanced and Loud, and you can switch between them.
Change it
Say “make the title bigger” in the same chat, or edit the words and numbers by hand for free.
Or start from a design that is already made.
The top row was made by the design director from one-line requests. The two rows below are the Design pack: 31 hand-made graphics that are free and work offline, with titles, maps, charts, lists, quotes and logo stings.
29 chart types
Every number has to trace back to your request, script or data.
Maps
True routes for a plane, car, train or ship, on a flat map or a globe.
Brand kits
Colours, fonts and logos in one kit. Import them from your website.
Skills
More than a hundred craft guides the AI reads, and you can write your own.
AI Director
Ask the AI to edit. You approve each step.
Direct my video reads the whole transcript, asks a few quick questions, then shows you a plan with a price on every line. Nothing runs until you press Approve plan.
It reads the take, then asks.
Up to four quick questions about what you want, plus how the graphics should be made.
AI replies scripted for this demo. Prices are from a run on an OpenRouter key. On a ChatGPT or Claude plan, the text steps show no charge.
Or ask in the chat.
Type what you want in your own words. Changes arrive as cards you apply, and a checkpoint is saved before each message so you can go back.
Tighten it
Removed 4 filler words and shortened 2 long pauses. One step in Undo.
Example replies. Attach images, spreadsheets, PDFs or Word files for it to read, and read or change the prompts it uses.
Turn one long video into short clips.
Ask for “5 short clips”. The AI finds the moments, scores each from 0 to 100, and makes each one its own composition. Frame the speakers with AI keeps the frame on whoever is talking, with Follow, Split, Wide or Screen and face. Or frame a clip by hand: drag a window over the picture, fixed or keyframed.
16:9 as filmed
9:16, stays on her
0:00
AI replies scripted for this demo. On Windows a clip follows the pointer or uses a still crop: following a face in short clips is Mac only.
Step 3 · Export
Export the file. Post it anywhere.
Cutscript writes the finished video to your computer, in the shape you choose: 16:9, 9:16, 1:1 or 4:5. You post it yourself: Cutscript does not upload anything for you today.
Try it. Pick a shape, then press Export. This demo exports nothing.
Export
A video with zooms, a camera bubble or graphics exports as MP4 H.264. It never scales a video up.
M4A, WAV, MP3 or FLAC
Mono or stereo, 44.1 or 48 kHz
No watermark on audio exports
Final Cut Pro, Premiere, DaVinci Resolve or an EDLPro
Your cuts and audio clips go across
Zooms, graphics and other motion effects stay behind, so export a video as well if you need the look
Subtitles: SRT or VTT, timed to your edit
Transcript: TXT, Markdown, HTML, Word or RTF
YouTube chapters: Direct my video writes a list to copy
In the app, the file is saved to your computer. Post it wherever you like.
Move a project
A project package opens on another Mac or on Windows. It holds the originals unless you leave them out, so share an exported video when what you hid is private.
On your computer
Export needs no AI and no internet.
Several at once
Export compositions and Export all clips make one MP4 for each.
One selection
Export only the words or the range you selected.
A warning, if needed
If a tracked blur lost its subject, the export tells you and jumps there.
Your AI
Use the ChatGPT or Claude plan you already pay for.
Cutscript sells no AI and adds nothing to the price. Sign in with your plan, and the AI chat, Direct my video, short clips and new graphics run on it. Cutscript shows no charge for them: they count against your plan’s own limits.
Try it. Tick what you already have.
What you can run
Recording, timeline, zooms, blur, Design pack, exportNo AI needed
AI chat, Direct my video, short clips, new graphics
Transcripts, so you can edit by text and add captions
Voiceover
Music and sound effects
AI pictures
Every AI feature, and what runs it.
A tick means that source can run the feature, and a row says so if it also relies on another row. The columns you ticked above light up, and so do the features they run.
Swipe sideways to see every column.
Your subscription runs text only, shows no charge in Cutscript, and counts against your plan’s own limits. Your key is billed by that provider, and Cutscript adds nothing to the price. Your computer is free, and works offline once a model is downloaded. Editing by text, captions, the AI chat and short clips all start from a transcript. Cutscript is not affiliated with these companies.
Plans cover the writing, not the media
Transcripts, voices, music and pictures need a key or a local model. You pay the provider directly.
You see the price first
Most paid steps show a typical and a maximum price. The Transcribe button, automatic transcripts and a project made from the home screen start without asking.
Set your own limits
Spending limits cap the AI director’s OpenRouter and fal.ai spend, per message, per day and per design. Transcripts, voices, music and sound effects show their price first but are not capped. Your keys stay encrypted on your computer.
Local models
Or run the AI on your own computer.
Download a model once. It costs nothing to run, works offline, and your audio and text stay on your computer. It is slower than the cloud, and small models make more mistakes. The cloud stays the default until you choose a local model.
Try it. How much memory does your computer have?
The AI director’s text
Ollama Mac and Windows
We would pickQwen3.5 9B7.6 GB
You install Ollama yourself, from ollama.com. It covers text only.
Transcripts
Whisper Apple silicon Mac
We would pickWhisper large-v3 turbo629 MB
99 languages (Distil-Whisper is English only), and no speaker labels. This is what lets you edit by text for free.
Voices
On your Mac Apple silicon Mac
We would pickQwen3-TTS 0.6B1.1 GB
For voiceovers and line-by-line dubs. Qwen3-TTS speaks 10 languages. Pocket TTS can copy a voice, with that person’s permission.
What stays on your computer
Cutscript sends no analytics or crash reports. Your recordings stay on your computer unless you run an AI step, and then only what that step needs goes to the provider you chose.
These need no model to download and no AI account
Filler, pause and repeat clean-up, once you have a transcript
Finding faces
Blur all faces and emails Mac
Remove background Pro
Tracking, zooms and cursor effects
Built-in voice clean-up
The Design pack
Export
Compare
A recorder and an editor, for less than most charge for one.
Cutscript against Descript, Screen Studio, Camtasia, CapCut, OpusClip and Loom, row by row.
Swipe sideways to see every column.
Prices in US dollars before tax, from each company’s own pricing page on 21 October 2026. Where a company shows two prices for a plan, the range is given. A question mark means we could not find it in their published docs, not that it is missing. CapCut figures come from search summaries of its official pages, which blocked our checker, and a CapCut sale ran 20 to 30 October. Names are trademarks of their owners; Cutscript is not affiliated with them. Seen something out of date? Tell us in the forum.
Pricing
Free to edit. Pro to export without the watermark.
Free
$0
Free for as long as you like
Every feature except timeline export and Remove background
A free Cutscript account
Exports with a watermark
No watermark on audio, subtitle or transcript exports
Neither plan includes AI credits. AI runs on your own plan, your own keys or your own computer. See every feature on both plans.
Questions
Plain answers.
What does the AI cost?
It depends on what you use. A ChatGPT or Claude plan covers the AI chat, Direct my video, short clips and new graphics at no charge in Cutscript. A transcript comes from ElevenLabs, fal.ai or OpenRouter, and you pay them directly: Fish Transcribe 1 on OpenRouter, for example, is $0.36 per hour of audio. On an Apple silicon Mac, Whisper makes transcripts for free. Filler clean-up, zooms, the Design pack and export never cost anything.
Can I use it without any AI?
Yes. Recording, the timeline, zooms, cursor effects, blur, the Design pack and export need no AI. To edit by text and add captions you need a transcript, which comes from a provider or from Whisper on an Apple silicon Mac.
Does it post to YouTube or TikTok for me?
No. Cutscript exports a file to your computer, and you upload it yourself. Direct my video can write a YouTube chapter list for you to copy.
Is my video uploaded anywhere?
Not by Cutscript. Recording, editing and export happen on your computer. When you run a cloud AI step, what it needs goes straight from your computer to the provider you chose: for a transcript that is the audio only, not the video. With local models, nothing leaves.
What is Mac only?
Area recording, iPhone and iPad recording, sound from one app, drawing while you record, click tracking and the zooms and click effects that come from it, Blur all faces, Blur all emails, Apple Vision tracking, the Claude plan, local Whisper and local voices (these two also need Apple silicon), Apple Voice Isolation and taking the background out of a picture. On Windows the pointer is part of the recorded picture. Known limitations lists every difference.
Which computers does it run on?
Mac: Apple silicon or Intel, macOS 14 to open the app and macOS 15 to record. Windows: 64-bit Windows 10 (version 1809 or later) or Windows 11, from the Microsoft Store. Linux is in testing and is not ready to download yet.
Do I need an account?
Yes, a free one. Sign in once. The app checks your plan about once a day when you are online, and works offline for up to 14 days.
Is it finished?
It is a beta, version 0.1. Things will change, and we say so plainly in the changelog. iPhone and iPad recording and sound from one app are marked experimental.
Can I cancel Pro?
Yes, any time. Pro lasts to the end of the period you paid for. After that, new video exports have a watermark, timeline export and Remove background need Pro, and your projects stay on your computer.
