# About Viditor

Canonical URL: https://viditor.app/about
Last updated: 2026-10-13

> TL;DR: Viditor is an AI video editor for Mac that edits by chat, watches the actual frames, and runs on your existing Claude or ChatGPT plan. It's built by Ajay Pawriya, a designer who builds. We think AI should watch the video, footage should stay on your Mac, and you shouldn't need a second AI subscription. It's open to everyone.

Image: Illustrated Viditor card reading AI that sees every frame., with a filmstrip of a skateboarder above a timeline and one frame highlighted at the playhead.

Viditor is an AI video editor for Mac that edits by chat, watches the actual frames, and runs on your existing Claude or ChatGPT plan. You say the edit, it makes it on a real timeline, and you finish it by hand when you want to. This page is about who makes it, why, and what we won't compromise on.

## Who builds Viditor?

I'm Ajay Pawriya, a designer who builds. I design and develop Viditor and edit my own videos in it. Being the first user is the point: if an edit is slow or a result looks wrong, I run into it before you do.

We write in plain words, and "we" on this site means the people making Viditor.

## Why does Viditor exist?

Because short video has become part of everyone's job, and editing it is still slow. Founders need launch videos for features they shipped this afternoon. Agencies spend their days on revision rounds. Creators record a good take and then lose an hour to cutting pauses and placing captions. Most of that work is mechanical, and it's exactly the kind of work an AI agent should be able to do, if it can see what it's editing.

Existing tools tend to fall into two camps. Traditional editors give you full control and leave all the work to you. Transcript-based editors are fast at cutting words, but they edit a document of what was said, so they can't tell where your face is, whether a caption covers your mouth, or whether a shot froze. We wanted the speed of chat with the judgment of someone watching the picture, on a real timeline you can still edit by hand.

## What does Viditor believe?

Three beliefs shape every decision.

### AI should watch the video

An editor that only reads the transcript is guessing about the picture. Viditor measures faces and on-screen text in your footage, checks every edit against the composed frame, and lets the assistant look at still frames before it decides. Quality checks run after every batch of edits, and a critic can review the cut before you export. When a caption would sit on a face or text would run off the frame, the assistant hears about it and fixes it.

### Footage stays on your Mac

Editing, transcription and export run locally. Transcription uses Whisper on your Mac, so audio never leaves it. To edit, the assistant sees the transcript and the still frames it inspects, sent to Claude or ChatGPT under your own plan. Finished videos stay on your Mac until you upload them yourself. We say exactly what the beta collects on the [privacy page](https://viditor.app/privacy).

### An editor, not a generator

Viditor edits footage you recorded: it cuts, captions, adds music and builds text, chart and shape graphics from templates. It doesn’t generate video, create avatars or put words in anyone’s mouth. The AI is an editing assistant: it watches your frames and decides what to cut, where captions sit and what moves when. The only generative extra is optional and runs on your own account: sound effects made with your ElevenLabs key. Voiceover is your own voice, recorded with the mic.

### No second subscription

Most people who'd use an AI video editor already pay for Claude or ChatGPT. Viditor runs on that plan, through Claude Code or Codex, using their existing sign-in. We don't resell AI credits or meter your edits. The one exception is optional generated sound effects, which use your own ElevenLabs key because ElevenLabs bills for them separately.

## What else do we care about?

A few more principles show up throughout the app:

- **Every AI change is a real edit.** The assistant uses the same timeline, layers and keyframes you do. Everything it makes stays editable, and each request undoes in one step.
- **You approve what matters.** Plan mode lets you approve big edits first. Viditor never posts for you: the finished video goes out only when you upload it.
- **Say what's missing.** We'd rather list a limit than have you find it. Timelines max out at 5 minutes, export is 1080p at 30 fps, there are no speed ramps, and brand looks are per project. Each page on this site states the limits that matter for it.
- **Keyboard first, minimal chrome.** Common actions have shortcuts, and advanced controls stay folded until you need them.

## What can Viditor do today?

Viditor covers the path from raw footage to a finished MP4. It has a real timeline with layers, keyframes, motion scenes, word-anchored captions, music and sound effects, with stock media included. Comments pin to frames and **Fix with AI** sends them to the assistant. It removes silences and filler words, frames for 9:16, 1:1 or 16:9 with a safe-zone overlay, and reviews each export for loudness, peaks, black or frozen frames and long silences. The Exports tab lists everything you've exported, ready to play or show in Finder, and you post it from the platform's app or your own scheduler. See [chat video editing](https://viditor.app/features/chat-video-editing) for how editing by chat works.

## How is Viditor built?

Viditor is a desktop Mac app with a local editing engine. Preview and export use the same compositor, so what you watch is what you export. Every edit, by you or the assistant, goes through one validated engine with atomic saves and grouped undo. Transcription uses Whisper with word-level timing. Motion graphics are HTML scenes rendered frame by frame in a sandbox with the network blocked, then cached, so they look identical every time. The assistant reaches your project only through a fixed set of editing tools, and its checks run on the Mac.

## Where is Viditor now?

Viditor is in beta for Apple Silicon Macs on macOS 14 or later, open to everyone, sold as a one-time licence (see [pricing](https://viditor.app/pricing)). Updates install automatically, and feedback sent from the app comes straight to us. Release notes are on the [changelog](https://viditor.app/changelog).

## How do I get in touch?

Email [hello@viditor.app](mailto:hello@viditor.app) for anything: questions, partnerships or press. Journalists can find descriptions, facts and images on the [press page](https://viditor.app/press). To try Viditor, [download it](https://viditor.app/download).

## Frequently asked questions

### Who makes Viditor?

Viditor is built by Ajay Pawriya, a designer who builds. Ajay designs and develops the app, and edits videos in it too.

### Is Viditor available to everyone?

Yes, on Apple Silicon Macs with macOS 14 or later. Download it, sign in with your email and start editing.

### How much does Viditor cost?

Viditor is a one-time purchase: one payment for a licence on two Macs, with 7-day refunds. Chat editing runs on a Claude or ChatGPT plan you already pay for, so there's no separate AI charge from us.

### How do I contact the Viditor team?

Email hello@viditor.app. You can also send feedback from inside the app with Help → Send Feedback.

### Does Viditor upload my videos?

No. Your footage stays on your Mac and isn't uploaded to us. During the beta we store the text of AI conversations, with images removed, to find where the assistant goes wrong; the privacy page explains exactly what we collect.

## Related

- [Viditor press kit](https://viditor.app/press)
- [Edit video by chatting](https://viditor.app/features/chat-video-editing)
- [Private, local video editing with AI](https://viditor.app/features/private-local-editing)
- [Can Claude edit videos?](https://viditor.app/claude-video-editor)
- [Changelog](https://viditor.app/changelog)
- [Privacy](https://viditor.app/privacy)

Download Viditor for Mac (one-time licence, see https://viditor.app/pricing): https://viditor.app/download