Find the moment
Transcribe, score and group the strongest moments before you spend time editing.
AI video editing you can review
Turn long-form video into clips worth sharing.
ToClips finds the moments that carry a story, gives you the timestamps and reasoning, then helps you shape each clip for the channel it belongs on.
Podcast, interview, webinar and course clips · Agent workspace with EndTo accountKeep the context
Every candidate keeps its source timecode, transcript and selection reason beside the preview.
Ship the right format
Reframe to 9:16, 1:1 or 16:9, add captions, and export a clean package.
See it on a real recording
A 4 minute 45 second talk went in. Three vertical clips came out, each a complete idea with burned-in captions, an SRT/VTT file and a 3:4 cover. Every clip lists the exact range of the source it was cut from, so you can check the edit instead of trusting it.
Original script, programmatic visuals and synthetic Chinese speech, produced offline for this demo. All videos contain Chinese speech and captions.
Play and download every fileEverything between the long recording and the post
ToClips is an AI video clipper built around one job: turning speech-led long-form video into short clips you can publish. Each step below is a real part of the workspace, not a roadmap item.
Find the highlights in a long video
The Agent transcribes every source, looks for moments with a complete setup and payoff, and ranks them. Instead of a black-box cut, you get the reasoning.
- Start and end timecode for each clip
- Selection reason and confidence in the result manifest
- Works on podcasts, interviews, webinars and lessons
Turn several long videos into a batch of shorts
Upload up to 20 source files to one job, describe what you want in a sentence, and choose how many clips you need, anywhere from 1 to 200.
- One brief for the whole batch (up to 4,000 characters)
- Per-file upload progress and retry
- Download clip by clip or as one zip
Captions that follow the words
Every source gets a timecoded transcript, and every clip gets its own subtitle file. Burn captions into the video in one of three styles, or keep them as separate files.
- SRT and VTT for sources and clips
- Clean, bold or karaoke burned-in styles
- Chinese, English or automatic language detection
Every aspect ratio from one source
Pick one or more output ratios per batch. Each version is rendered at full resolution, for example 1080×1920 for 9:16, scaled to fill the frame and centre-cropped.
- 9:16, 3:4, 4:5, 1:1 and 16:9
- Platform presets for YouTube Shorts, Douyin and Xiaohongshu
- H.264/AAC MP4 output
A cover and post copy for every clip
Each clip comes with a cover image matched to the platform preset and an Agent-written title and caption, so a batch is ready to schedule, not just ready to watch.
- A JPG cover per clip: 3:4 (1080×1440) for Xiaohongshu, 9:16 for Douyin and Shorts
- Title, caption and hashtags per clip
- Everything listed in video-result.json
Fine-tune a single clip in the browser
When one clip needs a closer look, switch to refine mode: load the source, move the in and out points, correct a name in the captions, and render again. Drafts are saved with the project.
- Trim boundaries by hand
- Edit caption text before rendering
- Track render status and download the result
A clear path from recording to release.
The workspace keeps media, decisions and rendered files in one reviewable place.
Upload several sources
Add up to 20 videos or audio files to one batch job.
Describe it in one line
Say what you want, e.g. "cut 100 viral clips for Xiaohongshu".
Set the batch size
Choose the target number of clips, a platform preset and aspect ratios.
Get clips and covers
Download each clip with captions and a 3:4 Xiaohongshu cover, or the whole batch as a zip.
Every clip has a reason to exist.
ToClips makes the editorial decision visible. Preview every clip once the batch is done; the timecode, score and reason behind each one are written to the result manifest in the zip.
- ✓Timecoded transcript for every source (SRT / VTT)
- ✓Start and end timecode for each clip, in the result manifest
- ✓Selection reason, confidence and render status
CANDIDATES / REVIEW LISTTRANSCRIPT + TIME CODE
The one idea people rememberA complete thought with a clear payoff.
96%The surprising counterpointStrong contrast and a natural question hook.
91%A practical next stepSpecific advice that works without context.
87%What goes in, what comes out
The limits are the ones the workspace actually enforces, so you know before uploading whether a job will fit.
- Video sources
- MP4, MOV, WebM, M4V, MKVUp to 5 GB per file
- Audio sources
- MP3, WAV, M4A, AAC, FLAC, OGGUsed for transcripts; clips render from a video track
- Per job
- Up to 20 sources · 1–200 clips
- Output ratios
- 9:16 · 3:4 · 4:5 · 1:1 · 16:9Full resolution, centre-cropped
- Clips
- MP4 (H.264/AAC)Optional burned-in captions
- Subtitles
- SRT · VTTPer source and per clip
- Covers
- JPG, one per clip3:4 for Xiaohongshu; 9:16 for Douyin and Shorts; otherwise your first output ratio
- Package
- ZIP + video-result.json manifest
Who ToClips is for
It works best on video where the value is in what people say.
Podcasters
Turn each episode into a set of quotable vertical clips that keep the question and the answer together.
See the workflowCourse creators and webinar hosts
Pull teachable moments out of long sessions for follow-up emails, social posts and lesson previews.
See the workflowYouTube creators
Cut a long upload into a batch of Shorts, each with its own title and caption, without scrubbing the timeline.
See the workflowXiaohongshu and Douyin operators
Use the platform presets, per-clip covers and copy to fill a posting calendar from a few recordings.
Founders and interviewers
Share the strongest minutes of a talk, panel or customer interview with the source range attached for review.
Marketing and content teams
Hand a reviewer a zip with every clip, subtitle, cover and manifest instead of a folder of unnamed exports.
ToClips compared with editing by hand
A traditional editor is still the right tool for effects-heavy work. For speech-led clips, most of the time goes into the steps ToClips automates.
| Step | Editing by hand | With ToClips |
|---|---|---|
| Finding the moments | Watch or scrub the whole recording | Ranked candidates with timecodes and reasons |
| Captions | Transcribe, time and style each clip | SRT/VTT per clip, burned in if you want |
| Several platforms | Re-frame and export each ratio separately | Choose ratios once for the whole batch |
| Covers and copy | Design and write them one by one | A cover, title and caption per clip |
| Review | Check exports against the source manually | Source range for every clip in the manifest |
Where ToClips is not the best fit
Being clear about the limits saves you a wasted upload.
- Reframing is a centre crop. There is no automatic speaker tracking yet, so shots where the speaker sits at the edge of the frame may need a manual edit.
- Moment selection relies on the transcript. Music videos, gameplay or b-roll without speech will not produce meaningful highlights.
- Batch mode does not let you edit the transcript. To fix a name or term, open the clip in the browser editor before rendering.
- There is no link import. Upload the video file itself; ToClips does not download from YouTube or other platforms.
- It is not a multi-track editor. If you need effects, music beds or motion graphics, export the clip and finish it in your editor of choice.
- You need the rights to the media you upload, and you should review clips and captions before publishing.
Pay for the clips you make
ToClips uses EndTo credits. When you submit a job, the credits it may need are reserved; when it finishes they are settled against actual usage, and anything unused is returned.
The sample, the guides and the tool pages are free and need no account.
See credit packsYour footage stays yours
Uploaded media, transcripts, tasks and rendered files are stored with your EndTo account so you can come back to a project and keep editing.
Private projects are never published as public pages, and the workspace is only available after you sign in.
Privacy policyFrequently asked questions
What is ToClips?
ToClips is an AI video clipper from EndTo.AI. It turns long, speech-led videos such as podcasts, interviews, webinars and lessons into short captioned clips in the aspect ratios you choose, with a cover, title and caption for each clip.
What kind of video works best?
Recordings where people speak with a clear structure: podcasts, interviews, talks, panels, webinars and courses. Clips are chosen from the transcript, so footage without speech does not produce useful highlights.
How many clips can one job produce?
One batch job takes up to 20 source files and can produce anywhere from 1 to 200 clips. You set the target number when you submit the job.
Which file formats can I upload?
Video in MP4, MOV, WebM, M4V or MKV, and audio in MP3, WAV, M4A, AAC, FLAC or OGG, up to 5 GB per file. Clips are rendered from a video track, so upload the video recording when you want publishable clips.
Can I paste a YouTube link?
No. ToClips works on files you upload and does not download from YouTube or other platforms. Only upload video you have the right to use.
Which aspect ratios and platforms are supported?
You can render 9:16, 3:4, 4:5, 1:1 and 16:9 versions at full resolution in the same batch, with presets for YouTube Shorts, Douyin, Xiaohongshu or general use.
Do the clips come with captions?
Yes. Each clip gets an SRT and a VTT file, and you can burn captions into the video in a clean, bold or karaoke style. Transcription supports Chinese and English, or automatic detection.
Can I edit a clip after it is generated?
Yes. Open the clip in the browser editor to move its start and end points or correct the caption text, then render it again. Drafts are saved with the project.
How much does it cost?
ToClips uses EndTo credits. Credits are reserved when you submit a job, settled against actual usage when it finishes, and unused credits are returned. The sample and guides are free.
Do I need to install anything?
No. ToClips runs in the browser. Sign in with an EndTo account to open the workspace, upload media and download results.
Guides and related pages
Start with the recording you already have.
Watch the sample to see exactly what a finished batch contains, then upload your own long video when you are ready.