Hand over the recording.
Get clips ready to post.
Snipri finds the moments worth posting, reframes them to vertical with the speaker tracked, and burns in word-by-word captions.
90 free minutes · no card · sign in with Google first
- Upload a file
- Dropbox
- Google Drive
Most recordings come back in about 10 minutes, and the length of the video barely changes that. Your clips land in your library behind your login — Snipri never posts anything anywhere.
One recording in. A week of posts out.
A long conversation gives up the handful of moments that stand on their own. Each one comes back cut, reframed to 9:16 with the speaker tracked, and captioned word by word — a 1080×1920 file you can post to Shorts, TikTok or Reels as it is.
Five real clips out of one 29-minute interview, shown exactly as they came back. The recording is not ours — it is third-party footage, used to show what the renderer does, with no affiliation or endorsement. Snipri gives you the file; posting it is still yours.
Three steps, and you do one of them.
Hand over the video
Upload a file, or pick one out of Dropbox or Google Drive. That is the entire setup — no timeline, no editor to learn.
Snipri reads all of it
The audio is transcribed with word-level timings, then scored for moments that stand on their own. Weak ones are discarded rather than padded out.
Clips arrive ready to post
Vertical, speaker-tracked, captioned word by word, with an optional title card over the opening seconds. Restyle any of them later without re-running the job.
The weak moments never reach you.
Every candidate is scored one to ten on five dimensions. Hook and payoff are a hard gate — if either lands under seven, the moment is not delivered. Some videos give up ten clips. Some give up two. Padding the list would be the easy lie.
Sample scores from one delivered clip — the notch marks 7/10.
“Everyone tells you to niche down”
A strong opening that never lands the point. It stops at the gate, and you never have to watch it to find that out.
Captions
Every clip arrives captioned. Then every field is yours.
Clips are born with the karaoke look — heavy uppercase blocks, the spoken word travelling green. Every field underneath is yours: font, colour, position, motion, emoji, and a title card you write yourself. Restyling re-renders the clip without touching the source.
Choose a preset to get started.
Set caption length for readability.
A demo, and not a screenshot of one: the panel is the editor’s own controls and the caption is its own preview renderer — the one the burn is written against — playing a pass through the presets and the fields. The footage under it is that clip’s own render with the captions off. Third-party footage, no affiliation.
Trim it, or split it in two
Scrub the clip against a timeline of its own footage, pull the ends in, or cut it into two posts. Neither half goes back to the source video.
Your own text and images on top
Type a line over the clip or drop in a logo — up to 8 text layers and 4 images, each dragged where you want it and given its own window on the timeline.
An emoji on the word you choose
Tag a word and the emoji is composited into the render above the captions, arriving with a movement of its own.
A title card over the opening
The clip's own title, or a line you write instead, held on a card for 1 to 10 seconds and dragged and sized where you want it.
A brand every clip is born with
Save your look, logo, standing line, colours and your own uploaded font once. Every clip after that comes back already wearing them.
Restyling spends no minutes
Each change re-renders from the clip's caption-less base rather than from your video, so a look you keep changing costs you nothing but the changing.
Six ready-made looks, one click apart.
A look is the whole caption specified at once — the face, how the spoken word is marked, the colour, and the way the words arrive. The panel above draws all six; here are four of them on four real clips, each carrying a different face and a different colour. Those are fields inside a look, not four separate products.
And these are finished files. Each is one of those styles with its face and its colour changed — exactly what the panel above changes. All four recordings are third-party footage, used with no affiliation or endorsement.
Smart reframe
Framed on the person talking, automatically.
Snipri tracks the speaker and keeps them in frame, so a landscape podcast becomes a clean vertical clip instead of a centre crop that cuts someone in half. When a second person holds the frame it stacks them; when the source is narrower than 9:16 it pads with a blurred fill.
Input 16:9
Output 9:16
Output is 1080×1920 on every plan. The window over the source is the crop this clip was really made from. Third-party footage, no affiliation.
- Tracks the active speaker across the cut
- Stacks a two-shot when both people hold the frame
- Smoothed motion, so the crop never jitters
Minutes, not your whole afternoon.
| Task | With Snipri | By hand |
|---|---|---|
| Time per batch of clips | Minutes | Hours |
| Finding the best moments | Scored and gated | Scrubbing manually |
| Captions | Burned in, word by word | Typed by hand |
| Resizing to 9:16 | Speaker-tracked crop | Crop every clip |
Pricing
Priced by what you put in, not what you take out.
You are billed for minutes of source video processed. Cutting one hour into twelve clips costs exactly what cutting it into three costs — so there is never a reason to hold back on the thing the product exists for.
Free
90 minutes of source videoone-time allowance
- Word-level captions you can restyle
- Clips never expire
- Vertical 9:16, sized for TikTok, Reels and Shorts
- Watermark on output
Starter
200 minutes of source videoevery month
- No watermark
- Word-level captions you can restyle
- Uploads up to 3 GB
- Email support
Pro
500 minutes of source videoevery month
- No watermark
- Uploads up to 5 GB
- Priority email support
- Word-level captions you can restyle
Scale
1,500 minutes of source videoevery month
- Everything in Pro
- API access (coming soon)
- Dedicated support
Clips never expire on any plan. Paid plans renew monthly and can be cancelled from the receipt email LemonSqueezy sends with every charge.
Questions
No. Hand over the video, look at what comes back, and post it. There is no timeline to learn — the cutting, framing and captions are already done when the clips arrive.
About 10 minutes for most recordings, from handing the file over to the clips being in your library. The length of the video is not what costs the time — transcribing and choosing the moments barely change between a ten-minute recording and a half-hour one, and the rendering scales with how many clips come back rather than how long the source was. While a job runs, the Studio shows an estimate that re-works itself at each stage instead of counting down from one guess.
Nothing is published. Clips land in your library behind your login, and Snipri never posts on your behalf or puts them anywhere else — where a clip goes after you download it is entirely your call. The third parties that run the pipeline are listed in the privacy policy.
Every candidate moment is scored one to ten on five dimensions: hook, payoff, standalone, arc and quotability. Hook and payoff are a hard gate — if either is under seven the moment is not delivered, which is why some videos give up ten clips and some give up two. You can see all five scores, and the reason a moment was chosen, on any clip in your library.
Upload a file, or pick one out of Dropbox or Google Drive. There is no link field right now — the link most people reach for is a YouTube one, YouTube blocks downloads from datacentre servers, and a field whose commonest use is a refusal is worse than no field at all. Downloading from YouTube Studio and uploading the file works fine.
Only on the free tier. Every paid plan — Starter included — exports without one.
Vertical 1080×1920 clips between 15 and 60 seconds, up to ten per job, with word-level captions burned in and an optional title card over the opening seconds. You can restyle the captions on any clip afterwards without re-running the job.
English today. More are planned but none has a date, and we would rather say that than imply a language works before it does.
Early access
Be one of the first creators on Snipri.
Snipri is brand new and built by one person, which is why the feedback loop is short. 90 minutes of source video are free — enough for a full episode — and no card is needed to find out whether it works on your footage.
90 free minutes · no card · sign in with Google first