Turn one long video into
ready-to-post clips.
Upload a podcast, interview, stream, or talk. Larkveo finds the strongest standalone hooks, tracks the active speaker in 9:16, burns in word-timed captions, and generates carousel card packs.
“The biggest moon is Ganymede”
Turn the same video transcript into a 5–10 slide LinkedIn & Instagram carousel deck with custom brand themes.
We fed Larkveo a 10-minute interview.
Here’s what came back.
We uploaded “Is there life on Jupiter's moons?” by The Royal Society and requested 3 clips of about 30 seconds with Karaoke captions. Left is the original 16:9 frame; right is the finished 9:16 clip Larkveo cut automatically.
Framing note: Watch the last six seconds. The original cuts to a wide shot, but the crop stays where her face was, so you mostly see the coffee table. Each clip gets one fixed crop.
Word-timed transcript · Line by line
WhisperX Alignment- 0:00So the biggest moon is Ganymede.
- 0:04And it is about the same size as Mercury.
- 0:07So it's about a third of the size of the Earth.
- 0:13And if there is this liquid water ocean on Ganymede, which we're almost certain there is, if it's a global ocean, there'll be more water in that ocean than there is on the water oceans on the Earth.
- 0:26Wow.
- 0:27So that gives you an idea about how much water there will be out there.
- 0:32Yes.
Every spoken word receives millisecond start and end timestamps so on-screen captions highlight the active word in sync with the speaker’s voice.
AI Metadata & Social Copy
- Title
- The Moon with More Water Than Earth
- Hook score
- 91 / 100
- Why this moment
- It shares a mind-blowing scale comparison about a moon having more water than all of Earth's oceans combined.
- Post caption
- This moon is only a third of the size of Earth, but it might hold more water than all of our oceans combined.
- Hashtags
- #ganymede #jupiter #spacefacts #astronomy
All 3 clips extracted from this run · Ranked by hook score
1080×1920 MP4 + Cover JPG + SRT
Clip #1 · 36s94Will We Find Life on Europa?
8:49 → 9:25
Clip #2 · 32s91The Moon with More Water Than Earth
6:30 → 7:02 · Featured above
Clip #3 · 28s85Where to Find Water in Space
1:39 → 2:07
Source video: “Is there life on Jupiter's moons?” by The Royal Society, licensed under CC BY 3.0. Processed automatically by Larkveo.
One video in.
Every social format out.
AI Video Clipper & Face Reframing
Finds self-contained hooks in 2-hour recordings, scores each moment, and locks a steady 9:16 crop around the active speaker.
Kinetic Animated Captions
9 creator-tested subtitle styles burned in at 60fps clarity, timed to the millisecond and placed safely above platform UI chrome.
Carousel Card Packs
Turn any video or topic into a 5–10 slide LinkedIn & Instagram carousel deck with editorial typography and brand themes.
Master-grade exports ready for every feed.
Every clip comes out as a finished H.264 MP4 with captions already burned in, paired with an SRT file, a custom cover frame, and social copy.
- Clips per run
- 1 to 5 vertical clips per run — you pick the count before processing starts.
- Target duration
- 15s, 30s, 60s, or Auto (sizes each clip naturally to its thought, 15–60s).
- Aspect ratios
- 9:16 vertical (TikTok, Reels, Shorts) plus 4:5 and 1:1 for feed posts & carousels.
- Resolution
- Full 1080p HD exports: 1080×1920, 1080×1350, or 1080×1080.
- Video & audio
- MP4 container with H.264 video and crisp AAC stereo audio.
- Captions
- Burned-in word-by-word kinetic subtitles in 9 presets, plus downloadable SRT & TXT.
- Speaker framing
- Automated computer-vision face detection centered on the primary speaker.
- Bonus deliverables
- Custom cover image with hook headline, social post copy, hashtags, and carousel decks.
From raw recording to
social-ready assets in minutes.
No timeline scrubbing, keyframing, or manual subtitle typing. Here is every stage Larkveo runs behind the scenes when you upload a video.
- 01 / 06Ingest
Drop a video or paste a link
Upload an MP4, MOV, M4V, WEBM, or MKV file (up to 4 GB and 2 hours) or paste a public video link.
Is there life on Jupiter's moons?
10:21 · 1920×1080
- 02 / 06Transcribe
Word-accurate speech alignment
Larkveo transcribes every spoken word with millisecond start and end timestamps and detects the language automatically.
- 6:30So the biggest moon is Ganymede.
- 6:34And it is about the same size as Mercury.
- 6:38So it's about a third of the size of the Earth.
- 03 / 06Score
AI finds standalone hooks
Our model reads the full conversation to locate self-contained moments that open with a strong hook and land a complete thought.
#1#2#30:0010:21 - 04 / 06Reframe
Active speaker 9:16 crop
Computer vision locates the primary speaker in each segment and locks a steady 9:16 vertical frame around their face.
- 05 / 06Render
Burn in kinetic captions
Word-by-word animated captions are rendered directly into a crisp 1080×1920 MP4 alongside a custom cover image.
MP4 + Cover + Deck1080×1920 verticalReady to download - 06 / 06Publish
Download clips & carousel decks
Preview every clip, tweak titles or social copy, generate a multi-slide carousel card pack, and export in one click.
Export bundleZIP + Individualclip_01.mp4cover_01.jpgcaptions.srtcarousel_deck.zip
Captions engineered
for vertical screens.
Every word lights up at the exact millisecond it is spoken. Positioned in the safe zone above TikTok, Reels, and Shorts UI chrome so your hook never gets covered by platform buttons.
A clean workspace where
every clip lands ranked.
When a run finishes, your clips appear in the Larkveo dashboard sorted by hook score. Open any clip to preview playback, edit the title and social caption, or export the MP4, SRT, and cover frame.
Hours of timeline work,
compressed into one run.
Larkveo doesn’t replace your creative judgment — it eliminates the repetitive mechanical steps between finishing a recording and publishing short-form clips.
What Larkveo doesn’t do yet.
We prefer to tell you our exact engineering trade-offs upfront rather than let you discover them halfway through an upload.
Some AI picks will miss.
Moments are chosen by an AI reading the transcript. Occasionally it starts a beat late or stops before the punchline lands — always preview before posting.
Fixed crop per clip.
Each clip locks a steady vertical crop around the primary face. If two speakers sit far apart in a wide shot, one speaker may sit outside the 9:16 frame.
Built for spoken-word video.
Moments are selected from speech. Recordings without dialogue fall back to evenly spaced segments.
Captions can mishear rare words.
Unusual names, slang, and technical terms are the main trouble spots. Inline word editing inside Larkveo is on the roadmap.
Optimized for Latin & Cyrillic scripts.
Our burned-in caption fonts support English, Spanish, French, German, Portuguese, Russian, and more. CJK and RTL scripts are still in progress.
Direct downloads — you control publishing.
Larkveo never posts to your social accounts without you. You download the 1080×1920 MP4s and carousel ZIPs and publish on your schedule.
Free during the public beta.
No credit card required.
Minutes are counted by source video duration — a 30-minute podcast uses 30 minutes of your monthly quota. Failed or canceled runs never count against you.
Full access to the Larkveo video clipping pipeline, kinetic captions, and carousel card packs while we finish our studio editor.
- 120 minutes of video processing every month
- Up to 5 vertical 1080×1920 clips per run
- Source videos up to 2 hours or 4 GB
- 2 concurrent video runs processing at once
- All 9 kinetic caption presets (9:16, 4:5, 1:1)
- AI hook score, title, post copy & hashtags
- Carousel Card Packs for LinkedIn & Instagram
- Zero watermarks on any exported MP4 or slide
For daily podcasters, agencies, and creator teams running high-volume episodes every week.
- 600+ monthly video processing minutes
- Up to 15 clips extracted per long-form episode
- Inline transcript & word-level caption editor
- Custom brand fonts, colors & logo watermarks
- Priority GPU rendering queue (3× faster)
- Team shared workspaces & bulk ZIP export
Questions, answered plainly.
How long does a video run take?
Which video formats and links can I use?
What happens to my uploaded recordings and data?
Do I need the rights to the videos I upload?
What is included in the free public beta?
Can I edit a clip after Larkveo renders it?
What are Carousel Card Packs?
“I built Larkveo so the best moments inside long conversations don’t stay buried inside an hour-long timeline.”
Got a long recording?
Turn it into clips today.
Drop a podcast, interview, or talk into Larkveo and see which standalone moments it pulls out in one run.