ChatGPT Video Editor in 2026: What the New CapCut ChatGPT Plugin Will Do, and When a Highlighter Edit Is Faster

Short answer: On 24 October 2026, at CapCutWorldLA, CapCut announced a plugin “coming soon” to ChatGPT that brings its editing and generation tools into the chat. No launch date was given. Until it ships, the closest thing is CapCut’s existing CapCut × Codex workflow in the ChatGPT desktop app.

If you came here because you searched “ChatGPT video editor” (about 480 searches a month in the US, per DataForSEO on 25 October), the useful question is not whether a chat can edit video. It is which kind of edit goes faster when you describe it in words, and which goes faster when you just point at the lines you want to keep. We make CutFast, a browser editor built around the second idea, so read the comparison with that in mind; we say where CapCut is the better pick.

What CapCut actually announced

CapCut’s official post says the plugin “brings CapCut’s video editing and generation tools alongside ChatGPT”: start with an idea in ChatGPT, turn it into a video with the plugin inside the chat, then continue in CapCut. The Another Daily AI Newsletter of 25 October reports the same and notes there is no date yet.

It is the third time CapCut and ChatGPT have met. A 2023 CapCut plugin in the old ChatGPT plugin store turned a one-line prompt into a short video with script, voiceover and stock footage (Search Engine Journal, July 2023); that plugin store is gone. The current CapCut × Codex page runs in the ChatGPT desktop app.

	2023 plugin	CapCut × Codex (today)	ChatGPT plugin (announced)
Where	Old ChatGPT plugin store	ChatGPT desktop app	ChatGPT
Input	One-line prompt	Your clips or a draft + chat instructions	An idea or your footage
Output	Generated short video	Editable rough cut in CapCut	Edited / generated video
Status	Retired	Available; varies by account and region	“Coming soon”, no date

Practical rule: Plan around what you can open today. “Coming soon” has no date, and the version that ships may differ from the keynote.

What chat-based editing can already do

CapCut’s Codex page lists what the chat workflow handles once your media is in: generate images or video from a description, or upload your own draft; adjust the timeline by asking (shorten scenes, reorder clips, replace visuals, change text timing, update music); build captions and caption styles; add transitions, filters and effects; run a “smart rough-cut” to arrange clips into a first draft; and open the full CapCut editor for track-level work.

That is a good fit for assembled video: B-roll, product clips, music, titles, templates. The instructions are about objects you can name (“the second clip”, “the intro music”) and the result is a draft you then polish.

The page is also clear about limits. It is for the ChatGPT desktop app, not a browser-only tool. Availability “may vary” by app version, account and region. And generated results are a starting point to review, not a finished video.

The video below shows what editing with ChatGPT looks like in October 2026, so you can judge the chat workflow for yourself.

Video: Kevin Stratvert · How to Edit Videos with ChatGPT (October 2026)

Practical rule: Chat editing shines when you can name what to change. If you can only describe it as “the part where I ramble”, you need a different handle.

Where describing cuts in words breaks down

Talking-head video, podcasts, tutorials and interviews are cut by sentence, not by clip. The edit you want sounds like “keep the second answer, drop the false start, drop the ‘um, so, basically’ before the demo”. Saying that in chat is possible, but you end up quoting the transcript back to the assistant, then checking every cut point by ear.

Three things make it slow:

1. You still have to read the transcript. The assistant cannot know which take you prefer. You find it, then describe it.
2. Word-level precision. Filler words and half-second silences live between words. “Remove the pauses” is easy; “remove the pause after ‘pricing’ but keep the one before the punchline” takes several rounds.
3. Review loops. Each instruction produces a new draft to scrub through. Five small fixes become five previews.

Illustration: describing cuts in chat (top) vs selecting the lines to keep (bottom)

Practical rule: If the thing you would type into chat is a quote from the transcript, skip the chat and select the quote.

The highlighter method: edit by selecting lines

CutFast skips the description step. You open a local file or paste a link (YouTube, Bilibili, TikTok, Xiaohongshu, Xiaoyuzhou podcasts), the AI transcribes it, and the transcript becomes the timeline:

1. Pre-transcription. Speech is turned into subtitle lines. Filler words, repeats and silences are marked automatically.
2. Highlight to keep. Click or drag over the lines you want, like a highlighter. You are selecting text, not dragging a playhead.
3. Export. The highlighted segments export as one clip at original quality. The homepage puts the whole loop at about five minutes.

Local files are processed in the browser and are not uploaded. For podcasts or long recordings where you only want the dead air gone, the silence remover does that one job without a project.

Screenshot: CutFast · highlight subtitle lines to keep; filler and silence are marked for removal

Which one to use for which video

Your video	Better fit	Why
Product promo, B-roll montage, music-driven short	CapCut (Codex today, ChatGPT plugin later)	Templates, effects, stock assets; edits are about clips you can name
Generate footage from an idea	CapCut plugin	Generation is part of CapCut’s pitch; CutFast does not generate footage
Talking-head, tutorial, interview, podcast clip	CutFast	Cut by sentence; filler and silence handled at word level
Sensitive recordings (client calls, internal demos)	CutFast	Local files stay on the device
Need a full multi-track timeline	CapCut	CutFast is deliberately not a full NLE

They also chain well: rough-cut the spoken part in CutFast, then bring the cleaned clip into CapCut for music, titles and effects.

Practical rule: Pick the tool by what your video is made of. Mostly speech → cut by text. Mostly visuals → cut by clips.

Cost and privacy, side by side

CapCut’s 2026 price change moved the full AI toolkit to the $19.99/month Pro tier (Newsweek); we broke the tiers down in our CapCut price-hike guide. The ChatGPT plugin’s pricing has not been announced. CapCut’s Codex page says you can start the workflow for free, without listing paid tiers.

CutFast pricing is on the homepage: Free, 3 edits a day; pay-as-you-go at $30; lifetime at $399. Processing for local files happens in your browser, so there is no upload queue and the footage does not leave your machine.

FAQ

Is the CapCut ChatGPT plugin available now? Not as of 25 October 2026. CapCut announced it at CapCutWorldLA on 24 October as “coming soon” with no date. The CapCut × Codex workflow in the ChatGPT desktop app is available today, depending on account and region.

Can ChatGPT edit videos by itself? Not without a tool. It needs a plugin such as CapCut’s to touch a timeline. The chat writes the instructions; the plugin does the cutting.

Does CutFast work inside ChatGPT? No. CutFast runs in your browser (and as a desktop app). You can use both: clean the spoken part in CutFast, then finish in CapCut.

What is the best ChatGPT video editor for talking-head videos? For speech-heavy video, selecting transcript lines is faster than describing cuts in chat. Try that first, then move to chat editing for music and effects.

Is CutFast free? Yes, 3 edits a day on the free plan, no install needed for browser editing.

Try the highlighter edit

If your next video is mostly someone talking, open it in CutFast and highlight the lines you want before you write a single chat instruction: edit a video with CutFast.

CutFast Team