Coming to Mac and WindowsJoin the waitlist

Say what you want. Saycut makes it.

Bring a script, a voice recording or a video, then tell your AI which one you need next: a sales letter, a social clip, a talking-head video. Each is its own build, and you approve the story first.

What we changed this monthPlaceholderFounder update

PlaceholderOnline courses

PlaceholderOffer page

Booked out is not the same as paid wellPlaceholderPodcast clips

PlaceholderConsulting

Your first call should not be your pitchPlaceholderCoaching

Say it once, then say it betterPlaceholderPersonal brand

PlaceholderLaunch teaser

Here is what I tell every clientPlaceholderWellness coaching

Same take, fewer umsPlaceholderWorkshop recording

Why our cohort stays smallPlaceholderGroup programs

PlaceholderProduct launch

Doors open FridayPlaceholderEvents

Sample videos coming. These show the shapes and looks.

Start with a script, a voice or a video.Saycut makes the rest.

No recording? Your AI works from your script, and you add a voice: a built-in one, or your own ElevenLabs or Cartesia account.

Home

Example data

Your scriptWord for word

Your voiceA recording

Your videoRaw take, 4:12

Example look
Sales letterPortrait 4:5
Example look
Kinetic textTall 9:16
Example look
Social clipTall 9:16
Example look
Talking head + graphicsSquare 1:1
Clean up raw videoWide 16:9

Swipe to see all five

The five kinds on the Home screen. You ask for each one, and your AI builds itIllustrative product animation

Two ways to pay for your AI.Neither one goes through Saycut.

Saycut doesn't charge for AI. Builds, takes and fixes use your own AI plan's limits or your API key's billing.

Recommended for most people

Flat monthly subscription

Use the AI plan you already pay for. Your usage counts against that plan's limits.

You chat in your AI company's own app, and it drives Saycut.

For people who prefer to pay by usage

Your own API key

An API key is a code from your AI company. They bill you for what you use.

It stays in your agent tool, such as Claude Code or Codex, and Saycut never runs AI on it.

Works with the AI app you already use, and whichever model you pick in it. Switch apps any time: your projects stay in Saycut.

Claude

Codex

Gemini

Grok

Cursor Any AI tool

Work where you like.Your AI app or Saycut, same video.

You connect one AI to make a video. After that, ask it for changes or make small ones by hand.

From your AI app or code editor

Ask Claude, Codex, Cursor or another AI tool. It builds and changes the video through Saycut.

Your AI app

Turn my script into a 3-minute sales letter, tall and wide.

The story outline is ready in Saycut. Approve it and I'll start building.

For developers: it connects over MCP or the CLI.

Inside Saycut

Watch the live preview, fix small things by hand on a light timeline, and leave a note for your AI. Then export.

Saycut's light timeline with scenes, words, music, sound effects and overlays

Example data

The timeline is for tweaks, not a full editor. The "Send to your AI" box passes your request to your AI app.

Same video, four shapes,each laid out for its shape.

Tick the sizes you want on the Export screen, and you get a file for each one.

Export

Example data

Wide 16:9coaching-offer-vsl_16x9.mp4

Frame rebuilt from the Export thumbnail, with a stock clip as a placeholderIllustrative product animation

5kinds of video

4sizes: Wide, Tall, Square, Portrait

2desktop platforms: Mac and Windows

Pick a kind of video,and get the exact words for your AI.

Placeholder
Sales letter
Slides and a voice that sell your offer.
One ideaYour offer page and your launch emails.
Placeholder
Kinetic text
Big moving words with real energy.
One ideaA launch teaser, or the first ten seconds of an ad.
The clips you post between launchesPlaceholder
Social clip
Short, vertical, captioned word by word.
One ideaThe clips you post between launches.
What the numbers sayPlaceholder
Talking head + graphics
You on camera, with numbers and charts.
One ideaCourse lessons and client case studies.
Placeholder
Clean up raw video
Cut the ums, pauses and retakes.
One ideaWorkshop recordings and other long footage.

You approve the storybefore your AI builds a single scene.

Read the shape of the video and approve it. "Show me 3 more" gets new takes for any part.

Set up screen with a video preview and the plan for the video

Clean up footage screen: 4:12 before, 3:36 after, with suggested cuts on the waveform

Story outline screen: eight parts from Hook to Close and an Approve and start building button

Live build screen: scenes filling in while the AI asks which price the call to action should show

Preview and tweak screen. Beside it, someone asks their AI app to make the 3× bigger, the AI replies that it changed scene 3, and the 3× in the video grows

Export screen with four sizes, quality and the list of files to be made

Story outlineExample data

Tweak it after it's built,by hand or by telling your AI.

Example data

Story outline

Go through it part by part. "Show me 3 more" gives you new takes for any part before you approve.

samehour,3×

AllBold socialKaraokeOne big wordPills & boxesClean subtitle

Claude's pickBold pillsPills & boxes

Example data

Captions

Try a style on every scene. SRT and VTT files come with your export.

Example data

Clean up

Your AI marks the ums, pauses and retakes. Your original file is never changed.

Timeline with scenes, words, music, sound effects and a playhead

Example data

Preview & tweak

See it live with no render wait. The light timeline handles the small changes and is not a full editor.

Alex RiveraBusiness coach

samehour,3×reachExample data

Brand kits

Titles, lower thirds, captions and end cards all use the kit a project picks.

Size cards for Wide, Tall and Square on the Export screen

Example data

Export sizes

Choose any of the four sizes and get a file for each, plus captions and suggested post text.

Your footage stays on your computer,and here is what your AI sees.

Footage stays off Saycut's serversFootage and voice recordings never go there, unless you use Saycut Cloud. The servers hold your account, devices, plan and usage counts.
Works without internetCompose, preview and export offline. Sign-in, stock search and outside services still need it.
Saycut never stores your AI loginIt never reads another app's credential files, and an API key stays in your agent tool.
A cost estimate comes firstVoices, AI presenters and AI shots show one before a paid media job. Each project's budget starts at zero.

Will it run on my computer?Here's what we expect it to need.

Proposed requirements for Saycut
Proposed requirements	Minimum	Recommended
Mac	macOS 13 or later, Apple silicon	Apple silicon, macOS 14 or later
Windows	Windows 10 22H2 or 11, 64-bit. Windows on Arm is not supported at launch	Windows 11
Memory	8 GB, for 1080p and videos up to about 10 minutes	16 GB or more, for 4K and long footage
Graphics	A working hardware H.264 encoder for 4K. Without one, exports cap at 1080p	Apple silicon, or a recent Intel, NVIDIA or AMD GPU
Disk	About 0.6 to 1.2 GB for the app, plus optional downloads and project space	An SSD with about twice your source footage free

These are proposed and still being tested. The first time you open Saycut, it encodes a real test clip and tells you plainly what your computer can export, such as "Your Mac can export up to 4K." See the full list

Questions

Saycut doesn't charge for AI. Builds, takes and fixes use your own AI plan's limits or your API key's billing.

You need one AI connected to make a video. The Recommended way is your AI app's monthly plan, and its limits apply. The Advanced way is your own API key in an agent tool, billed by your AI company.

Claude, Codex, Gemini, Grok, Cursor and other agent tools. On a ChatGPT plan you connect through the Codex app. You need one AI connected to make a video, and you can look around and open the sample video without one.

Saycut runs no AI of its own and has no built-in script writer. Your own AI does the thinking, and you approve the story outline before it builds.

Saycut stops at finished files on your computer. Each export can include a captions file and suggested post text, and you post them yourself.

It has a light timeline for tweaks. If you want to cut frame by frame all day, keep your editor. Saycut is for people who would rather say what they want and then fix the last few things.

English at launch, across the app, transcription, captions and voices.

Your AI drives Saycut over MCP or the CLI. The app detects supported clients on its Connect screen, and you can add your own agent.

We haven't set pricing yet. Join the waitlist and you'll hear when Saycut opens.