Find any moment in a video just by describing it

Ctrl+F for video: describe a scene, action, or line in plain words — AI watches the video, finds matches with precise timestamps, and cuts ready-to-use clips. It understands visuals, speech, sound, and on-screen text.

the part where the product is shown a person picks up a knife a line about a discount or price someone laughing text appears on screen
Starting…
Usually takes 1–4 minutes, depending on video length. Just keep this page open.

Video subtitles

Highlight the words you want right in the transcript — click individual words or drag across a phrase. You can select several parts in different places; they stay selected until you click “Clear selection.” Tap a line's timestamp to select the whole line. Each continuous selection becomes a separate segment. Then hit “Cut selection” and merge the segments into one video.
Highlight words or phrases in the subtitles above

How it works

1

Video + query

Paste a YouTube, Instagram, TikTok, RuTube, Yandex, or Vimeo link, or upload a file, and describe the moment in plain words.

2

AI watches the video

The AI analyzes the footage, listens to speech and sound, and reads on-screen text to find every match for your query.

3

Ready-to-use clips

Each moment comes with timestamps, a description, and a match confidence score. Download clips individually or as a single ZIP.

FAQ

How does searching a video by description work?

Paste a link or upload a file, then describe the moment in plain words — for example, “a person picks up a knife” or “the scene where the product is shown.” The AI watches the footage, listens to speech, reads on-screen text, finds matches, shows timestamps, and cuts ready-to-use clips.

Can I find a spoken phrase?

Yes. The AI analyzes the audio track: type the phrase or its gist, and the tool finds where it's said and shows the exact quote.

What videos are supported?

Links from YouTube, Instagram, TikTok, RuTube, Yandex Video, Vimeo, and more, plus MP4, MKV, AVI, MOV, and WEBM files up to 1 GB. Videos up to 60 minutes.

How many moments can it find?

Up to 10 of the most relevant per search. Each comes with timestamps, a confidence score, a scene description, and a ready-to-use clip.

Can the AI tell similar scenes apart — like a knife in a kitchen vs. in a fight?

Yes. The model understands context: setting, actions, speech, and sound. The more specific your query (“chopping vegetables with a knife in the kitchen”), the better the results.

People also use
Clip MakerTurn long videos into Reels and ShortsVideo VoiceoverAI voiceovers, no narrator neededYouTube SubtitlesExtract subtitles and transcripts
Found this tool useful?
RUEN
Sign in