ChatCut
ChatCut is a browser, desktop and agent-based AI video editor: describe an edit in plain English and it performs real cuts, captions, motion graphics, generated B-roll and music on a genuine multitrack timeline you can still adjust by hand.
What is ChatCut?
ChatCut is an AI video editor published by ChatCut Inc., an Austin company founded by two filmmakers. Its premise is that you should describe the edit rather than assemble it: you say what you want, the agent works out how, then it does the work on a real multitrack timeline. The distinction the company draws with rival AI editors is the heart of the product. Most of them render a template you cannot touch afterwards; here every instruction lands as ordinary clips, tracks and captions that you can still open and adjust by hand.
The agent watches your footage, works out what each clip is about, drops repeated takes, strips filler words and dead air, and assembles a first cut. Beside it sits text-based editing: every word in the transcript carries a timecode, so deleting a line removes the matching video and audio together. Transcription starts on upload with the language detected automatically, and the same transcript feeds word-level captions in more than 100 languages and over 20 styles you can save as templates.
Generation is folded into the same workspace rather than bolted onto it. Motion graphics, meaning chapter cards, lower thirds, animated charts, counters, timelines and logo reveals, are produced from a sentence and arrive as editable clips whose text, colours, fonts and timing change live with no re-render, and re-editing one costs nothing. Video generation runs on Seedance 2.5, Seedance 2.0 and Kling 3.0, image generation on GPT Image 2 and Nano Banana 2, and a royalty-free music generator writes a track to the exact length the cut needs. Voiceover, voice cloning, sound effects, LUTs and screen recording are documented too.
The product ships as three surfaces onto the same cloud project: a browser editor, a desktop application for macOS and Windows, and an agent plugin that lets Claude Code or Codex drive it over MCP. Exports cover MP4, WebM, MP3, SRT, ProRes 4444 with alpha, FCP7 XML for Premiere and DaVinci Resolve, and editable CapCut drafts. The site describes the product as currently in public beta.
What it does
- Turn a plain-English instruction into real cuts on a multitrack timeline
- Cut a video by editing its transcript, with video and audio following each word
- Strip filler words, false starts and dead air automatically, and find the best moments
- Generate editable motion graphics such as lower thirds, chapter cards, counters and animated charts from a sentence
- Produce word-level captions in more than 100 languages and over 20 styles
- Generate the shots you could not film, plus images, royalty-free music and voiceover, inside the project
- Drive the whole editor from Claude Code or Codex through the MCP agent plugin
When to use ChatCut / When not to
A quick filter to help you decide if ChatCut is the right fit.
When to use ChatCut
- Solo creators making talking-head videos for YouTube, who would rather record than spend the evening trimming filler words and dead air
- Social media and content managers turning one upload into vertical cuts, captioned highlights and platform-shaped clips for TikTok, Reels and Shorts
- Marketing and product teams producing demos, launch films and explainers that need lower thirds, animated charts and branded overlays without an After Effects licence
- Podcasters and interviewers who prefer to clean up a conversation by reading the transcript rather than scrubbing a waveform
- Educators and instructional designers cutting a long recorded session into chaptered lessons, with captions in more than a hundred languages
When not to use ChatCut
- Teams that need several editors working on the same timeline at once: real-time multi-user editing is not supported and is only listed as a roadmap item
- Developers looking to call video editing from their own software: ChatCut exposes an MCP plugin but publishes no API and no API documentation
- Organisations with a European compliance checklist: the site never mentions the GDPR, offers no data processing agreement, names no subprocessors and hosts everything in the United States
- Anyone working under sector-specific regulation, since the terms explicitly state the service is not tailored to HIPAA or FISMA and must not be used in breach of the GLBA
- Mobile-first editors, as there is no iOS or Android application, only a browser editor and macOS and Windows desktop builds
How to use ChatCut
A typical end-to-end flow, from setup to results.
- Pick your surface: the browser editor at app.chatcut.io, the desktop build for macOS or Windows, or the agent plugin
- Create a free account; no payment card is requested to start
- For desktop, download the binary matching your machine, offered for Apple Silicon, Intel Macs and Windows
- For the agent route, add the ChatCut marketplace and install the plugin, then authorise the account in the browser page it opens
- Check the install with the plugin list and the MCP server entry, then open a new session so the tools load
- Upload your footage into the media pool; transcription begins immediately and the language is detected on its own
- Describe the edit you want in the AI panel, referencing media where needed
- Watch the agent state its plan and apply each step, then review the result on the timeline
- Refine by prompting again, or open the inspector and change text, fonts, colours, sizes, positions and animation speed by hand
- Export to MP4, WebM, MP3, subtitles, transparent ProRes, XML for Premiere or DaVinci Resolve, or a CapCut draft
Pros & Cons
Pros
- Instructions produce genuine clips and tracks you can still edit afterwards, not a locked render
- The whole chain lives in one workspace: cutting, captions, motion graphics, generation and export
- Transcript-based editing makes talking-head cleanup fast, with no scrubbing
- Motion graphics are generated and stay editable without After Effects or keyframing, and re-editing them costs no credits
- The same project is reachable from a browser, a desktop app and a coding agent over MCP, which few video editors offer
- No watermark on exports, on any plan, and a genuinely usable free tier that asks for no payment card
- The terms commit explicitly to not training AI models on user media
Cons
- The product is openly labelled as being in public beta
- Real-time multi-user collaboration does not exist yet and is only promised on the roadmap
- No public API and no API documentation, so it cannot be integrated into another product
- No iOS or Android application; mobile users are left with the browser
- The compliance framework is thin: no GDPR mention, no DPA, no subprocessor list, no SOC 2 or ISO 27001 certification, and United States hosting with no regional choice
- Credit consumption varies with the AI model chosen, which makes the real cost of a finished video hard to anticipate in advance
- The free tier caps web exports at a cumulative sixty minutes, and transparent ProRes, local export and 4K are held back for paid or desktop use
Pricing & Plans
A permanent free plan is available and requires no payment card: it includes the full editor, uploads, transcription and manual timeline work, a one-off starting credit balance and a cumulative sixty-minute web export allowance. The lowest paid entry point is USD 25.00 per month for 100 credits per month, billed monthly and renewing automatically until cancelled. All amounts are charged in United States dollars.
- full editor
- uploads
- transcription and manual editing
- with a one-off starting credit balance (stated as 20 credits in the documentation and 15 in the homepage FAQ) and a cumulative 60-minute web export allowance
- no watermark
- USD 25 per month
- USD 45 per month
- listed against a USD 50 reference price
- USD 88 per month
- listed against a USD 100 reference price
- USD 160 per month
- listed against a USD 200 reference price
- Higher Pro tiers selectable at 1
- 200
- 2
- 000
- 3
- 000
- 4
- 000
- Every Pro tier bundles Seedance 2.5
- Seedance 2.0
- Kling 3.0
- GPT Image 2
- Nano Banana 2
- motion graphics
- AI music
- transparent motion-graphics export and all AI agent editing capabilities
- An annual billing toggle is offered on the pricing page
- though the annual amounts are not rendered in the static page
Data, GDPR & hosting
A consolidated view of how ChatCut handles your data.
GDPR overview
There is no mention of the GDPR anywhere on the site. The word does not appear once across the privacy policy, the terms, the usage policy or the documentation. The privacy policy, effective 7 June 2025, follows a United States template: it sets out no legal basis for processing, no right of access, rectification, objection or portability, no supervisory authority and no retention schedule. No Article 27 EU representative is designated and no data protection officer is named. No data processing agreement is published or offered, and no subprocessor list exists. The terms state that the service is hosted in the United States and that accessing it from another region counts as consent to that transfer. One concrete right is granted: account and data deletion on request by email, carried out within seven days. A complaints route is provided for California residents, with no European equivalent.
Who owns the data?
ChatCut Inc. owns the service itself, its source code, design, text and trademarks, but its terms carve out an explicit exception for what you make with it: the intellectual property clause states that it does not restrict your right to create, export and commercially publish your own video, audio, image or motion-graphics output, which is governed instead by the usage policy. Two asymmetries deserve attention. Anything posted to a public part of the service, such as a forum, grants the company a perpetual, irrevocable, worldwide, royalty-free licence for any purpose, and any suggestion or feedback sent to the company is assigned to it outright. On storage, the terms disclaim all liability for lost or corrupted user data.
Reuse rights
Users may reuse their own output without asking permission. The terms say so directly: the ownership clause protects the vendor's own website, software and branding, and does not restrict creating, exporting and commercially publishing the video, audio, images and motion graphics you produce. The limits are your own: you must hold the rights to the footage, music, logos and likenesses you upload or generate, and the usage policy forbids passing synthetic or altered media off as authentic. On the vendor's side, the terms open with a plain statement that user media is not used to train AI models. Personal data collected is limited to name, email and usage data, used to run the service, honour the contract, send offers and analyse campaigns, with cookies and web beacons administered in-house. No subprocessor is named anywhere on the site.
Data retention & training
Hosting summary
The terms of service state plainly that the service is hosted in the United States, and that accessing it from another region counts as consent to having your data transferred to and processed there. The privacy policy places the company in Texas, United States. No alternative region is offered and there is no data residency option, so European or other non-US customers have no way to keep processing local. No hosting provider or subprocessor is named anywhere on the site, and no hosting-related certification such as SOC 2 or ISO 27001 is published. As a separate network observation rather than a company statement, the domain resolves to a Cloudflare anycast address, which indicates a content delivery network in front of the site and says nothing about where data is actually stored. The terms also note that the service is not tailored to sector-specific regimes such as HIPAA or FISMA.
Things to keep in mind
Risks and trade-offs to weigh before adopting ChatCut.
- Synthetic media and consent: the usage policy forbids generating or altering a real person's face, body, voice, signature or identity without permission, and bans deepfakes, fabricated events and altered public statements. The tool makes this easy enough that the rule is worth reading before, not after
- Disclosure duty: when a reasonable viewer could be misled, the policy requires you to say the content is edited or AI-generated. Publishing a generated scene as documentary footage is an ethical and reputational risk the tool itself flags
- Erosion of editorial judgement: an agent that finds the highlights, cuts the takes and builds a narrative is making editorial choices on your behalf. The policy insists a human reviewer stay in the loop for news, politics, finance, health, legal or safety content, and that advice is easy to skip when the output already looks finished
- Rights on uploaded material: you are responsible for the footage, music, logos and third-party media you bring in, and the terms shift all liability for that onto you
- Broad licence on public contributions: anything posted to a public part of the service grants the company a perpetual, irrevocable, worldwide, royalty-free licence for any purpose
- Weak recourse: liability is capped at what you paid over the previous six months, disputes go to binding arbitration in Kent, Delaware with class actions excluded, and the company reserves the right to delete an account at any time without notice
- Data leaving your jurisdiction: everything is hosted in the United States and using the service from elsewhere counts as consent to that transfer, with no European framework, no processing agreement and no published retention schedule for active accounts
Setup & Integrations
Technical difficulty
Very low on the web: open the editor, create an account, upload footage. No installation and no payment card. Low on desktop: download the macOS or Windows build. Moderate for the agent route, which is the only demanding path: two terminal commands to add the marketplace and install the plugin, then an OAuth login in the browser. The marketplace clone is large and can take several minutes, and a recent Claude Code version is required. No editing experience is needed either way, since transcription starts automatically and the timeline opens only if you want it.
Deployment
Integrations
Supported languages
Behind ChatCut
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement ChatCut.
Frequently asked questions
What exactly is ChatCut?
Can I still edit by hand once the AI has worked?
Is there a free plan, and what does it include?
How much does the cheapest paid plan cost?
Where can I run it?
Will my footage be used to train AI models?
How many languages do the captions cover?
Can several people edit the same project at the same time?
Where is my data hosted and how do I delete it?
Is there a minimum age?
Should you pick ChatCut?
ChatCut occupies a genuinely distinct position among AI video tools. Rather than generating a finished clip you cannot touch, it puts an agent on top of a real non-linear editor and leaves the timeline open, so a prompt-driven first cut can still be reworked frame by frame. That single design decision, combined with transcript-based editing, generative motion graphics that stay editable, and captions in more than a hundred languages, makes it credible for creators, marketing teams, podcasters and educators who publish video regularly and do not want to learn a traditional editing suite.
The reservations are those of a young product, and the company states them itself: ChatCut is in public beta. Real-time multi-user editing does not exist yet, there is no public API and no mobile application, and the credit-based pricing makes the cost of a finished video hard to forecast since consumption depends on which generative model you call. The free tier is unusually honest for the category, with no watermark and no payment card, but it caps web exports at sixty minutes in total.
The area that deserves the most scrutiny is legal. The company commits clearly to not training on user media and to leaving commercial rights over your output alone, which are the two questions creators ask first. Beyond that, the framework is thin: no mention of the GDPR anywhere, no data processing agreement, no named subprocessors, no published security certification and United States hosting with no regional option. European organisations with a compliance checklist should treat that as a blocker rather than an oversight. Individual creators and small teams working in the United States will find far less to worry about, and a well-documented usage policy that is clearer than most about what synthetic media may and may not be used for.
- Choosing a selection results in a full page refresh.
- Opens in a new window.