All guides
Free to try

AI Video Editor

The AI Video Editor is an AgentReel tool that edits a video you already have from a plain-language instruction. You upload footage and type what should change: cut the pauses, keep the best 30 seconds, add captions, make it vertical. AI makes the edit, checks preview stills and renders a 1080p MP4. There is no timeline to learn.

This guide also covers three related video tools: the YouTube Clip Maker for YouTube clip links and free trims in your browser, the Talking-Head Cutter for pauses, filler words and retakes in a long recording, and Video Remix for a new script in the structure of a video that works.

Is it free? Edits and previews cost 0 credits: 3 edits a day before your first purchase and 30 a day after it. The final render needs purchased credits: 15 credits up to 30 seconds, 26 credits for one minute and 50 credits for the two-minute maximum. Free check-in credits do not pay this render.

Quick answers

What can I ask the editor to do?

Cuts, trims and removing pauses; captions and translated subtitles; vertical, square or wide framing; titles, a logo or a progress bar; color grades, black and white and vignettes; speed changes and slow motion; crossfades between clips; music under the speech.

How does the editor know what is in my video?

The speech is transcribed with a time on every word, so cuts land between words and captions match the voice. Your browser also makes a contact sheet of frames across the video, each marked with its time, so the editor sees the framing, the people and any text in the footage.

Can I edit the same video many times?

Yes. Ask for a change, look at the result and ask for the next one. The editor keeps everything you did not mention. Every round is saved as a version you can go back to.

Do I need editing or coding skills?

No. Type what you want in plain words, in English, Chinese or any other language. If you know HTML, CSS or GSAP, you can also open the code and change a value by hand.

What you can make

  • A 30 second vertical short from the best answer of a long interview, with bold captions.
  • A clean version of a recording with the pauses and false starts cut.
  • A square or vertical version of a wide video for social feeds.
  • A video with a title card, a logo and a progress bar.
  • A graded look, a black and white section or slow motion on one moment.

How it works

  1. Open the editor. Go to the AI Video Editor.
  2. Upload your files. Add a video up to 10 minutes long, plus music, a voice track or a logo if you want them. Up to 5 files per edit.
  3. Type the edit. Say the result you want, up to 1000 characters, for example "a 30 second vertical short with the best answer and bold captions".
  4. Check the preview. The real renderer takes 4 stills. The editor fixes a caption over a face, text cut off at the edge or a black gap before you see the result.
  5. Edit again. Ask for the next change as often as your daily limit allows.
  6. Render. Render the MP4 when it looks right. The button shows the price first. The video is saved in My Works.

Options and limits

ItemLimit / options
Video uploadMP4, MOV or WebM of any size, up to 10 minutes
Audio uploadMP3, M4A or WAV, any size
Image uploadPNG, JPEG or WebP, up to 10 MB
Files per editUp to 5
InstructionUp to 1000 characters per edit
Output1080p MP4 in 9:16, 16:9 or 1:1, up to 2 minutes
Edits and previews3 a day before your first purchase, 30 a day after it
VersionsEvery round is saved and can be restored
StorageUploads stay in a folder of your account under random names

Pricing

Rendered lengthCredits
Up to 30 seconds15
Up to 40 seconds18
Up to 50 seconds22
Up to 60 seconds26
Up to 2 minutes50

Writing the edit, the transcript and the previews cost 0 credits. The render uses purchased credits only. A failed render is refunded automatically. See Credits, plans and free usage for packs and plans.

Tips for better results

  • Say the result, not the steps. Name the length, the format and the platform if they matter.
  • Refer to the speech. "Start at the question about the weather" is precise, because the editor has the transcript.
  • Change one thing at a time. Small requests such as "make the captions bigger" keep the rest of the edit as it is.
  • Use Code to Video for a new video. The editor starts from your footage; Code to Video writes a video from a description.

FAQ

How long can the edited video be?

Up to 2 minutes. The source video can be up to 10 minutes long, so you can cut a short out of a longer recording.

Can free check-in credits pay the render?

No. The render of the AI Video Editor needs purchased credits. Free credits pay for Z-Image images and video templates, Code to Video included.

Can I add captions in another language?

Yes. Ask for captions in another language than the speech, and the editor translates them.

Where is my footage stored?

Your uploads are stored in a folder of your account under random names and are used only for your edits and renders.

What if a render fails?

The credits come back to your balance automatically, and you can render again.

Long video to Shorts

The Long video to Shorts mode of the AI Video Editor turns one long recording (a podcast, an interview, a lesson or a talking-head video) into several vertical Shorts at once. It transcribes the speech word by word, picks up to 8 moments that stand on their own, and renders each one as a 9:16 MP4 that follows the speaker, with word-by-word captions and a hook title.

Is it free? No. Long video to Shorts uses purchased credits; free check-in credits do not pay for it. Finding the clips costs about 15 credits for a 10 minute video and about 39 credits for a 60 minute video. Each rendered Short costs 2 credits. A failed render is refunded automatically.

How it works

  1. Open the mode. On the AI Video Editor, choose Long video to Shorts.
  2. Upload one video. MP4 or MOV of any size, up to 90 minutes. Pick the spoken language: English, Chinese, Spanish, Portuguese, French, German, Italian, Japanese or Korean.
  3. Wait for the clips. The tool transcribes the speech and ranks up to 8 clips of 15 to 60 seconds. Every clip starts and ends on a full sentence and shows its score, its grades for Hook, Flow, Value and Trend, its reason and a title.
  4. Adjust. Tick the clips you want (the top 5 are ticked), move the start or the end by one sentence, and edit the title. Choose the framing and the caption style. Press Command K (Control K on Windows) to search the clip titles and the whole transcript and jump to a moment.
  5. Render. Render the ticked clips. Each Short renders on its own, so one failure never stops the others.
  6. Download. Download each Short, or all of them in one ZIP file. Every Short is also saved in My Works.

Options

OptionChoices
FramingFollow the speaker (the crop tracks the face), or fit the whole frame over a blurred background (for screen recordings)
CaptionsWord-by-word highlight in a box, outline or yellow style, or no captions
Speaker colorsFor a recording with two or more voices: each speaker gets a caption color (the first speaker keeps the style color), in the preview and in the rendered clip
TitleThe hook title on the video, or none
ClipsUp to 8 candidates, 15 to 60 seconds each; Find more adds new candidates for a fee

A vertical source is cropped to 9:16 around the center. A change to a clip or an option renders only the clips it changes; an unchanged clip keeps its file and is not charged again.

YouTube Clip Maker and video trimmer

The YouTube Clip Maker does three jobs. It makes a link or an embed code that plays only one part of a YouTube video. It trims your own video files in your browser, with optional burned-in subtitles. And its AI search finds the best moments in your video for you. It suits creators, teachers, podcasters and social media managers who cut long videos into short clips.

Is it free? Clipping a YouTube link and trimming your own video are free and need no account. AI highlight search costs 15 credits per search, paid with purchased credits.

What you can make with the Clip Maker

  • A YouTube link that starts at the exact moment you want people to see.
  • An embed code for a blog or a website that plays only one part of a video, from your start to your end.
  • An MP4 clip cut from your own video file, ready to post on TikTok, Reels or Shorts.
  • A clip with subtitles drawn into the picture, so people who watch without sound can follow.
  • A short list of the best moments in a long recording, found by AI from your description.

Clip a YouTube link

  1. Open the Clip Maker and stay on the first tab, "Clip a YouTube link".
  2. Paste a YouTube link (a watch, youtu.be, Shorts, live or embed link) and press "Load video".
  3. Set the start and the end. "Use current time" takes the time from the player.
  4. Press "Preview clip" to check it.
  5. Copy the link that starts at your clip, or copy the embed code for just this part.

Trim a video file

  1. Open the second tab, "Trim a video file".
  2. Choose or drop an MP4, MOV, M4V, WebM or MKV file. The file stays on your device.
  3. Set the start and the end.
  4. Optional: add subtitles as an .srt or .vtt file. "Show the subtitles in the clip" is on by default; turn it off to keep the lines out of the picture.
  5. Press "Cut clip", then download the MP4.

Find clips with AI

  1. Load your own video file in the second tab. Add subtitles if you have them.
  2. Describe the moment, such as "the goal" or "when he says let's go". Leave the field empty to find the best moments.
  3. Press "Find clips with AI". The button shows the price.
  4. You get up to 5 suggested ranges, each with a match percentage. Press "Use" on the one you like, then cut it.

Clip Maker options and limits

ItemLimit / options
YouTube linksWatch, youtu.be, Shorts, live and embed links
YouTube outputA share link that starts at your start time, and an embed code that starts and stops at your times
Video files to trimMP4, MOV, M4V, WebM or MKV
Trim outputMP4
Subtitles.srt or .vtt, optional, burned into the clip
AI search descriptionUp to 200 characters, or empty for the best moments
AI search resultsUp to 5 ranges, each 3 to 60 seconds
What the AI search sendsSmall still frames and the subtitle text, never the video file
BrowsersChrome or Safari work best. If a file does not open, export it as H.264 MP4 first.

Clip Maker pricing

FeaturePrice
Clip a YouTube linkFree, no account needed
Trim your own video fileFree, no account needed
Burn in subtitlesFree
AI highlight search15 credits per search, purchased credits

The AI search button shows the price before you start. If a search fails, the credits are refunded automatically. Any credit purchase unlocks AI search. The Agent can run the same search on a video you attach, at the same price.

Clip Maker tips

  • Add subtitles before an AI search. The search reads the subtitle text, so it finds spoken moments much better.
  • Describe one clear moment in a few words. "The goal" works better than a long paragraph.
  • Use "Preview clip" before you copy a YouTube link or an embed code.
  • For a vertical post, trim the clip here, then cut several clips into one video with a script and a voice in the AI B-Roll Editor.
  • If the cut fails, try Chrome or Safari, or export the video as H.264 MP4 first.

Clip Maker questions

Does this tool download YouTube videos? No. It makes a link and an embed code that play part of the video on YouTube. To get a clip as a file, trim a video that you own in the second tab.

Is my video uploaded when I trim it? No. The cut runs in your browser, and nothing is uploaded. It is free. See privacy, rights and safe use.

Is my video uploaded for the AI search? No. Your browser sends small still frames and the subtitle text, never the video file itself. Trimming stays free and stays on your device.

Do I need an account? Not for clipping a YouTube link or for trimming a file. AI search needs an account with purchased credits.

What if the AI search finds nothing? The tool tells you when there is no clear match. Try other words, or add subtitles. If the search itself fails, the credits are refunded automatically.

Can I cut clips from someone else's video? Clip links and embed codes play the video on YouTube. Only trim and repost video files that you own or have permission to use.

Talking-Head Cutter

The Talking-Head Cutter makes a rough cut of a talking-to-camera recording. It transcribes your MP4 or MOV word by word, in one of 9 languages, and marks every long pause, filler word and abandoned take as a cut. You check the cuts in the transcript, then render one jump-cut MP4 or export an editable timeline for Premiere Pro, Final Cut Pro or DaVinci Resolve. Choose it over the AI Video Editor when you want to keep the whole talk: it takes a recording of up to 31 minutes, while the editor takes 10 minutes of footage and renders up to 2 minutes.

Is it free? No. The cutter uses purchased credits. A 10 minute recording costs about 7 credits to analyze and a 5 minute cut costs about 7 credits to render. Changing the cut, the preview and the timeline export cost 0 credits.

What the cutter removes

Pauses longer than your threshold, filler words such as um, uh and erm (and like or you know when they stand alone), and the earlier attempts of a sentence you said again. Each language has its own list of hesitation sounds and fillers, so a Chinese or a Spanish recording is cleaned in the same way. Only your last complete take stays. The cutter understands English, Chinese, Spanish, Portuguese, French, German, Italian, Japanese and Korean speech.

You can also choose what happens to swear words: keep them, cut them, mute them or bleep them. Every choice except keep also masks the word in the captions.

Every cut is struck through in the transcript. Click it to keep it, or click again to cut it. The preview updates at once, and nothing is charged until you render.

What you can make with the cutter

  • A tight jump cut of a YouTube talking-head video, without the dead air.
  • A clean course lesson or tutorial with the false starts removed.
  • A rough cut that you finish in Premiere Pro, Final Cut Pro or DaVinci Resolve.
  • A captioned version for social media, with a subtle punch-in that hides the jumps.
  • A multicam cut that switches between up to 4 cameras of the same take.
  • A list of YouTube chapters for the video description.

Cut a recording step by step

  1. Start a project. Open the Talking-Head Cutter and start a new project.
  2. Upload the recording. Drop in one MP4 or MOV with a sound track, straight from your camera or phone, and pick the spoken language. A large file is fine. Add up to 3 more camera files of the same take if you filmed with several cameras.
  3. Let AI find the cuts. The recording is transcribed with word timings, and pauses, fillers and retakes are marked.
  4. Review the transcript. Click any struck word or pause to keep it, and move the pause slider. Preview the result in the browser.
  5. Render or export. Render one jump-cut MP4 with optional captions and punch-in, or export the timeline. Copy the YouTube chapters if the editor lists them.
  6. Download. The cut is saved in My Works.

Cutter options and limits

ItemLimit / options
SourceOne MP4 or MOV per project, with a sound track
File sizeNo limit
LengthUp to 31 minutes
SpeechEnglish, Chinese, Spanish, Portuguese, French, German, Italian, Japanese or Korean
Pause threshold0.2 to 3 seconds, 0.5 seconds by default
Padding0 to 0.5 seconds kept around each word, 0.1 seconds by default
CutsLong pauses, filler words, earlier takes of a repeated sentence, words repeated after a false start inside a sentence; filler and retake cuts can be switched off
Swear wordsKeep, cut, mute or bleep; every choice except keep masks the word in the captions
CaptionsOptional, burned in, 12 styles: box, outline, yellow, shadow, banner, glow, bold, highlight, pop, one word, karaoke, reveal
Extra camerasUp to 3 more files of the same take, each up to 40 minutes, lined up by their sound; the cut switches to the camera of whoever speaks, and you can set each camera as a close-up or a wide shot, nudge its sync by one frame or choose the camera of a shot yourself
YouTube chaptersListed with a copy button when the cut has at least 3 chapters of 10 seconds or more, the first at 0:00
Punch-inOptional 1.08x zoom on every other cut, on by default
OutputMP4 in the shape of your recording, up to 1920 px on the long side, 30 fps, normalized sound
Timeline exportOne ZIP with FCPXML for Final Cut Pro and DaVinci Resolve, Final Cut Pro 7 XML for Premiere Pro, SRT captions, and a chapters file when the cut has chapters

Relink the exported timeline to your original recording and every cut is in place. Some phones record HEVC MOV files, which only Safari previews. The cut, the render and the export still work, and the rendered MP4 plays everywhere.

Cutter pricing

Each step is charged on its own, and a step that fails is refunded automatically.

StepTypical cost
Reading the recordingAbout 2 credits, once per recording
Transcript with word timingsAbout 2 credits for 5 minutes, 6 for 15 minutes, 12 for 31 minutes
Retake check and chapter titlesAbout 1 credit
Lining up extra cameras2 credits for 2 or 3 cameras, 3 credits for 4 cameras
RenderAbout 2 credits per 2 minutes of finished cut, plus 1 credit to join the parts
Changing cuts, preview, timeline export0

The cutter uses purchased credits only. Free check-in credits do not pay for it.

Cutter tips

  • Record with clean sound. Clear speech gives exact word timings and better cuts.
  • Say the whole sentence again. When you restart, repeat the full sentence, so the cutter can keep the last complete take.
  • Start with the default pause threshold. Move the slider only if the cut feels too tight or too loose.
  • Export when you want to polish. Use the timeline export to add B-roll and music in your own editor.

Cutter questions

What does the cutter make? One MP4 jump cut of your own recording, with every cut you approved removed and the sound normalized, plus optional burned-in captions.

Can I render again after a change? Yes. Change the cut and render again. Each render is charged on its own. If a render fails partway and you render the same cut again, the parts that already finished are reused, so you never pay for them twice.

What if a step fails? The credits for the failed step come back automatically. If the transcript fails, press "Retry analysis of this recording" to try the same recording again without a new upload.

Video Remix

Video Remix breaks down a short video you like and writes a new script in the same structure for your own topic. It works in two steps. First it breaks down the reference: the hook of the first 3 seconds, the beats with timestamps, every shot, the cuts per minute, the caption style, the sound and the call to action. Then it writes a new voiceover and a shot-by-shot storyboard. It reuses the structure only, never the words. The AI Video Editor changes footage you already have; Video Remix plans a new video before you shoot or make it.

Is it free? No. A breakdown costs 13 purchased credits and a remix costs 13 purchased credits. A step that fails is refunded. The finished video is made in another tool, which charges its own credits.

What you can make with Video Remix

  • A new video in the format of a competitor's best video, with your own words and product.
  • A stronger hook, built the same way as a hook that holds attention.
  • A shot list for an editor, with shot sizes, camera moves, on-screen text and timings.
  • A series of videos that keep the same rhythm from one episode to the next.

Remix a video step by step

  1. Upload a reference. In Video Remix, pick a Reel, Short, TikTok or ad of up to 3 minutes.
  2. Get the breakdown. AI reports the hook and why it works, the beats with timestamps, every shot, the cuts per minute, the captions, the sound and the call to action. 13 credits.
  3. Describe your topic. Write your topic or product, up to 2000 characters, and add notes about the audience, the tone or a line you must say, up to 1000 characters.
  4. Get the remix. AI writes a new voiceover and a shot-by-shot storyboard with the same structure and pace. 13 credits.
  5. Make the video. Send the result to Code to Video, the AI Video Ad Generator or the AI B-Roll Editor. The tool opens with your script filled in.

Video Remix options and limits

ItemLimit / options
ReferenceAn MP4, MOV, M4V or WebM file of up to 3 minutes
LinksYouTube and TikTok links are not accepted yet
What is sentStill frames and the audio track, which are not stored; the video file stays in your browser
TopicUp to 2000 characters
NotesUp to 1000 characters
BreakdownHook of the first 3 seconds, beats, shots, cuts per minute, captions, sound, call to action
RemixNew voiceover and shot-by-shot storyboard
LanguagesThe reference can be in any language that speech recognition understands; the report follows the site language and the script follows your topic
Saved resultsThe breakdown, the remix, your topic and your notes stay in this browser, so a reload keeps them

Video Remix pricing

StepCredits
Breakdown13
Remix13

Video Remix uses purchased credits only, and free check-in credits do not pay for it. A failed step is refunded automatically. The tool that makes the final video charges its own price. The Agent can also break down a video you attach, at the same price.

Video Remix tips

  • Pick a reference that works for your audience. The remix copies its structure, so start from a video with a strong hook and a clear order of beats.
  • Give real facts about your product. The remix uses only the facts you give.
  • Remix one breakdown many times. Use the same breakdown for several topics to plan a series.

Video Remix questions

Is my video uploaded? No. Your browser reads still frames and the audio track from the file. Only those frames and the audio are sent for the breakdown, and they are not stored.

Will the remix copy the original video? No. It keeps the skeleton of the reference and writes every line new for your topic. It never copies the reference's words, claims, brand or people, and it never invents facts about your product.

Can I paste a YouTube or TikTok link? Not yet. Upload a file that you have the rights to view.

How long can the reference video be? Up to 3 minutes, which covers Reels, Shorts, TikToks and most ads.

What if a step fails? The credits of the failed step come back automatically.