Clip API for long video
One long video in. Ranked shorts out.
Send a podcast, stream or webinar and get back its best moments as vertical clips: scored 0 to 100, reframed to 9:16 around whoever is talking, captioned word by word. $0.03 per source minute, every clip included.
- First $1 free
- Up to 20 clips per video
- MP4, SRT and VTT per clip
6 clips, ranked
done in 3 min 46 s
- 960:42
We were three weeks from shutting down
- 910:37
The pricing mistake that cost us a year
- 880:51
Why I stopped reading resumes
Call it from anything that speaks HTTPS
- n8n
- Make
- Zapier
- Python
- Node.js
- Workers
- Supabase
- Vercel
- Lovable
- Bolt
How it works
Everything an editor does, in one request
You send the recording. clipsdk does the listening, the picking, the cropping and the captions, then hands back finished clips.
- en58:06Daniel
28:33.4Honestly? We were
28:34.6three weeks from
28:35.9shutting the whole
1Transcribe
Every word gets a timestamp, so cuts land between words instead of in the middle of one.
- transcript
- start_s
- 96
We were three weeks from shutting down
Opens on the stakes, tells the whole story and lands the turn
91The pricing mistake that cost us a year
88Why I stopped reading resumes
2Rank
Each candidate moment is scored 0 to 100 for its hook, a complete thought and a payoff.
- score
- reason
Reframe
Faces are tracked and the 9:16 frame follows whoever is talking, cut by cut.
- aspect
- reframe
- clip-1.mp4
- clip-1.jpg
- clip-1.srt
- clip-1.vtt
4Caption and render
Captions are burned in word by word, and every clip comes back as MP4 with a poster, SRT and VTT.
- video
- srt
- vtt
Moment ranking
Ranked by what makes people stop scrolling
clipsdk reads the whole transcript and scores every candidate moment from 0 to 100: a hook in the first seconds, one complete thought, a clear payoff. Each clip comes back with its score and a one-line reason, so your app can post the 90s on its own and queue the rest for a human.
- 5 clips by default, up to 20, each 10 to 90 seconds
- Steer it with a prompt like "moments about pricing"
- A title and a ready-to-post caption for every clip
off-script-112.mp4
6 moments
We were three weeks from shutting down
Opens on the stakes, tells the whole story and lands the turn
28:33 to 29:15
Honestly? We were three weeks from shutting the whole thing down. Payroll was due Friday, the bank had said no twice, and then one cold email changed everything.
- 91
The pricing mistake that cost us a year
Counterintuitive claim up front with a concrete number
12:07
- 88
Why I stopped reading resumes
Clear opinion, easy to share, ends on a takeaway
41:20
- 84
The one number we check every morning
Specific and actionable, works without context
31:42
- 79
The cold email that saved the company
Strong story with a slower start
7:35
Speaker reframe
Reframed around whoever is talking
Face tracking finds every person in a 16:9 recording and keeps the one who is speaking in the vertical frame, cutting between them as the conversation moves. One speaker who stays put can use a fixed center crop, and slides or screen shares can keep the whole frame.
- 9:16, 1:1, 4:5 or 16:9 at 1080p
- Cuts on speaker changes, holds through short interruptions
- auto, center or none per request
tracking 2 speakers, following the voice
Animated captions
Captions for the sound-off crowd
Every word is timed to the frame and burned into the clip in the style you pick. Want to style them yourself? Turn captions off and take the SRT and VTT files that ship with every clip.
- Three presets: bold, karaoke and minimal
- Spoken language detected automatically
- SRT and VTT files with every clip
The API
One request. Every clip, ranked.
POST a video and get finished clips back in the same response, or have them delivered to your webhook. No queues to run, no GPUs to rent, no ffmpeg to babysit.
POSTapi.clipsdk.com/v1/clips
curl https://api.clipsdk.com/v1/clips \
-H "Authorization: Bearer $CLIPSDK_KEY" \
-H "Content-Type: application/json" \
-d '{
"video_url": "https://cdn.example.com/off-script-112.mp4",
"clips": 6,
"aspect": "9:16",
"captions": "bold",
"webhook_url": "https://example.com/hooks/clipsdk"
}'job_q8w2e4r6t1y3u5i7o9p0succeeded in 3 min 46 s
- 96
We were three weeks from shutting down
28:33 to 29:15
MP4SRTVTT - 91
The pricing mistake that cost us a year
12:07 to 12:45
MP4SRTVTT - 88
Why I stopped reading resumes
41:20 to 42:11
MP4SRTVTT - 84
The one number we check every morning
31:42 to 32:15
MP4SRTVTT - 79
The cold email that saved the company
7:35 to 8:21
MP4SRTVTT - 73
What I'd tell myself at 25
51:50 to 52:19
MP4SRTVTT
Wait, poll or webhook
Hold the request open for up to 300 seconds, poll the job, or get a signed POST the moment it finishes.
Any recording you own
Upload up to 5 GB through a presigned URL, or pass a direct link to an MP4, MOV, WebM, MKV file.
Files for 7 days
Each clip ships as MP4 with a poster frame, SRT and VTT, served with Range support for instant previews.
Automations
Plug it into n8n, Make or Zapier
A recording lands in a folder, clipsdk ranks the clips, the strongest go straight to YouTube and the rest wait for a human. Import the n8n workflow and it runs on its own.
- One HTTP node, no custom code
- A signed webhook resumes the workflow when clips are ready
- Scores turn the posting rule into a single IF
Google Drive
New file in /episodes
clipsdk
Create clips
IF
score is 85 or higher
YouTube
Upload as a Short
Airtable
Add to review queue
Use cases
If people talk in it, there are clips in it
Podcasts, streams, lessons and talks all go through the same call. Only the settings change.
- 60 min episode: ~$1.80
Podcasts
Every episode, a week of shorts
See the setup - 3 hour VOD: ~$5.40
Livestreams
Hours of VOD down to the highlights
See the setup - 45 min lesson: ~$1.35
Courses and webinars
Lessons cut into teaser clips
See the setup - 30 min keynote: ~$0.90
Talks and keynotes
Every session, clipped by lunch
See the setup
Pricing
$0.03 a minute of source video. Every clip included.
You pay for the length of the recording you send, not for the clips you get back. Ask for 3 clips or 20, the bill is the same.
| Price | |
|---|---|
| Per source minute | $0.03, every clip included |
| Past 2,000 minutes a month | $0.02 a minute |
| Free credit | $1, or 33 minutes |
| Billing | Prepaid credit, no subscription |
| Per video | |
| Longest video | 3 hours, 5 GB |
| File types | MP4, MOV, WebM, MKV |
| Clips per job | 1 to 20, 5 by default |
| Clip length | 10 to 90 seconds |
| Output sizes | 9:16 at 1080x19201:1 at 1080x10804:5 at 1080x135016:9 at 1920x1080 |
| Files per clip | MP4, poster, SRT and VTT |
20 h
- clipsdk~$36.00
- OpusClip~$116.40
over its 15 h a month API cap
- Klap~$177.60
Competitor prices are list prices per source minute as of October 2026: OpusClip $0.097, Klap $0.148.
FAQ
Questions, answered
What does clipsdk do?
It turns one long video into short vertical clips. Send a podcast, stream, webinar or talk and clipsdk transcribes it, scores every candidate moment from 0 to 100, reframes the best ones to 9:16 around whoever is speaking and burns in word-by-word captions. You get MP4s back with titles, timestamps, transcripts and SRT files.
How much does it cost?
$0.03 per minute of source video, and every clip cut from it is included. A 60 minute episode costs ~$1.80 whether you ask for 3 clips or 20. Your first $1 is free, there's no subscription, and past 2,000 minutes in a calendar month the rate drops to $0.02.
Which videos can I send?
MP4, MOV, WebM or MKV files up to 3 hours and 5 GB, uploaded to us or passed as a direct file URL. YouTube page links aren't accepted: send the file you own instead.
How does it pick the moments?
From the transcript. Every word is timed, then a ranking model looks for a hook in the opening seconds, a complete thought and a payoff, and gives each candidate a score and a one-line reason. You can steer it with a prompt like "moments about pricing" and set the clip length from 10 to 90 seconds.
How do I get the clips back?
Poll the job, hold the create call open for up to 300 seconds, or pass a webhook_url and we POST the finished job to you, signed with HMAC-SHA256. Clip files stay downloadable for 7 days.
Can I use my own caption style?
Pick one of three burned-in styles: bold, karaoke or minimal. For anything else, set captions to false and style them yourself from the SRT and VTT files that come with every clip.
Start free
Your back catalog is full of shorts
Your first $1 is free: 33 minutes of video, every clip included.
