0 uploads · 0 accounts · free

Drop a video.
Leave with the deck.

Video2Any watches your recording, catches every slide change, and hands you an editable PowerPoint plus subtitles. All of it happens inside your browser: your file never touches a server.

Live meeting or class? Record your screen instead.

Captured slide at 00:0000:00
Captured slide at 00:0400:04
Captured slide at 00:1100:11
Captured slide at 00:1700:17
Captured slide at 01:0401:04
A real run: this 80 second clip came in as 80 candidate frames and left as 10 slides.

It reads the video before it cuts it

Most converters grab a frame every N seconds and call it a day. Video2Any first scans your file, measures how much actually changes between frames, and picks a strategy to match:

MODE: bimodal

Slide decks

Long still stretches, sharp jumps at each transition. The threshold lands in the gap, so a one word edit on a white slide still registers.

MODE: motion

Talking heads & camera video

Everything moves, so naive detection keeps every frame. Video2Any raises the bar to typical motion plus two deviations and keeps only real scene changes.

MODE: static

One slide, an hour of audio

Compression noise never counts as a new slide. You get one clean frame, not four hundred duplicates of it.

The upload button is missing on purpose

Decoding, frame comparison, and the .pptx assembly all run in your browser. That is why the free tier has no daily quota, why an hour long lecture costs nothing to convert, and why a confidential recording stays confidential.

Where your file goes
Nowhere. It stays in the tab.
Server queue
None. Your machine, your pace.
Account required
No. Open the page, drop the file.

Subtitles from the same file

One more click runs Whisper speech recognition, also locally, and hands you a timestamped .srt. Pair it with the deck and every slide keeps the words that were said over it.

Try speech to SRT
3
00:00:11,200 --> 00:00:14,800
到达柜台身份证递给服务人员后

4
00:00:14,800 --> 00:00:17,300
这时柜台人员会给你一张登机牌

Or skip the browser: call the API

The same engine as a metered API, built for n8n, Make, and AI-agent tool calls: send a YouTube link, a Google Drive file, or an upload, and get timestamps, a deck, and subtitles back as JSON in one request.

Read the API reference
POST /v1/extract
{
  "video_url": "https://youtube.com/watch?v=...",
  "formats": "json,pptx"
}

The next slide you screenshot by hand is on you.

Convert a video now