Is anything sent to a server?

Nothing of yours. The application is static files with no API behind it, and the only later requests download the speech models from Hugging Face the first time a clip is transcribed.

Where is my project stored?

It lives in this browser’s storage, on this machine. Clearing the browser’s data for the site removes it, so keep a .cutvid backup of anything that matters.

Can I use it with the network off?

Yes, from the first visit on, which keeps a copy of the app in the browser. Transcription needs its model downloaded once first.

No. The source is not public at the moment. The claim above can be checked from the browser’s own network panel without it.

What are the limits on file size or length?

The machine and the browser’s storage quota set them, rather than a plan. There is no per-minute cap.

A video editor that never uploads the footage

cutvid is for material that cannot go to somebody else’s server: unreleased work, an NDA, a patient, a client, a case. The file is copied into the browser’s own storage, the render happens in the tab, and none of it is sent anywhere.

Start editing See everything it does

Edit video without uploading: what that means here, exactly

The phrase is used loosely; in cutvid it means there is no server to upload to. The application is a set of static files; once the page has loaded there is no API it talks to, no bucket the media is copied into, and no account the work is attached to.

The media: The browser reads it from disk and keeps it in its own storage, known by a content hash, so two projects can share one file without copying it.
The decode and the render: Both run in the tab on WebCodecs and WebAssembly, using the machine’s own hardware. An HEVC file is decoded by FFmpeg compiled to wasm.
The transcription: Whisper runs in the browser, and no audio is sent for it. The model is downloaded from Hugging Face the first time, and kept.
The export: It is encoded in the tab and written straight to disk.
The project: It is a CRDT document in the browser’s own storage, and it backs up as one .cutvid file with every original it uses.
Offline: The first visit keeps a copy of the app. With the network off everything above works, transcription once its model has been downloaded.

An offline video editor in the browser: how to check the claim rather than trust it

Open the network panel in the browser’s developer tools and cut something. After the page and its assets have loaded, editing makes no further requests: no upload, no telemetry, no analytics beacon. The only later downloads are the speech models, fetched from Hugging Face the first time a clip is transcribed; everything else, the text reader included, comes from this site. Or load the page once, turn the network off, and keep editing: the work carries on because there was never anything on the other end of it.

The projects screen: the projects on this device as cards, with the sort and the view

Who needs a video editor for confidential footage

Unreleased product footage and anything under embargo.
Work under an NDA, where a cloud editor is a contract problem rather than a preference.
Medical, legal and HR recordings, where the upload is the part that is not allowed.
Internal communications and anything with a customer’s data in frame.
Journalism and research with a source to protect.
Anywhere the network is slow, metered or absent, and a 4 GB upload is not going to happen.

It is a real editor, not a trimmer

Most tools that promise no upload are single-purpose trimmers. This one has a timeline with ripple trim and sync-locked tracks, nested sequences, keyframes on an anchor, 32 transitions, chroma key, LUT grading, titles that kern, a per-clip audio chain, transcription with speakers, captions, screen and camera recording, and an MP4 at the end.

What it does not do:

Export is H.264 in MP4, MOV or MKV, VP9 in WebM, or the audio alone as M4A, Opus or WAV. No ProRes, no DNxHD, no H.265 out, no image sequences.
No team features: no cloud project, no review link, no comments, no shared library. A project lives in one browser and moves as a file.
No stock library and nothing generative: no stock clips, no music bed, no AI b-roll, no generative fill.
No third-party plugins: no OpenFX, no VST, no motion-graphics templates from another app.
Color is LUTs and the effect rail, not scopes, curves and secondaries. A .cube grades a clip; there is no vectorscope to grade against.
It runs in Chrome, Edge and the other Chromium browsers, and in Firefox, where it draws with WebGL2. Safari is not checked yet.
It runs in a tab, on one machine. A long 4K timeline is heavier here than in a native application on the same computer.

Questions

Is anything sent to a server?: Nothing of yours. The application is static files with no API behind it, and the only later requests download the speech models from Hugging Face the first time a clip is transcribed.
Where is my project stored?: It lives in this browser’s storage, on this machine. Clearing the browser’s data for the site removes it, so keep a .cutvid backup of anything that matters.
Can I use it with the network off?: Yes, from the first visit on, which keeps a copy of the app in the browser. Transcription needs its model downloaded once first.
Is it open source?: No. The source is not public at the moment. The claim above can be checked from the browser’s own network panel without it.
What are the limits on file size or length?: The machine and the browser’s storage quota set them, rather than a plan. There is no per-minute cap.

Open it and cut something.

No account, no install, nothing uploaded.

Start editing See everything it does