How it works

From raw footage to viral clips in 3 steps

Upload once. Our AI pipeline takes care of transcription, virality scoring, reframing, and captioning. You just post.

Step 1·01

Upload your long-form video

Drag and drop your MP4 or MOV file (up to 4GB) onto the upload page. We accept podcasts, webinars, interviews, vlogs, lectures — any talking-head content works best.

Specs

FormatsMP4, MOV (QuickTime)
Max size4 GB
Best contentTalking-head, interview, podcast
Min length2 minutes recommended

What happens

  • File is uploaded directly to AWS S3 via a presigned URL — never touches our web server
  • Pre-flight normalization prepares the audio track for transcription
  • You receive a job ID and see real-time queue status
Step 2·02

AI pipeline processes your video

Our BullMQ worker running on AWS ECS Fargate picks up your job and runs the full AI pipeline — transcription, scoring, reframing, and clip extraction — in a single pass.

Specs

TranscriptionOpenAI Whisper
ScoringProprietary virality model
ReframingFace-tracking 9:16 crop
Avg timeUnder 3 minutes

What happens

  • Whisper transcribes every word with timestamp-level accuracy
  • Virality model scores each segment 0–100 based on hook strength, pace, and engagement signals
  • Smart reframing crops to 9:16 and tracks the main speaker face
  • Auto-captions are generated and synced to each clip
Step 3·03

Download and go viral

Your clips appear ranked by virality score. Download the ones you want, upload directly to YouTube, or follow the guided share flow for TikTok, Instagram, and Facebook.

Specs

DownloadPresigned S3 link (5-min expiry)
YouTubeDirect OAuth upload
TikTok / IG / FBGuided download-and-post
Clip formatMP4, H.264

What happens

  • Clips are sorted by virality score — highest first
  • Each clip shows duration, aspect ratio, and score badge
  • Connect YouTube once via OAuth and upload directly with title and privacy settings
  • For TikTok and Instagram, download the clip and the app opens automatically

The pipeline at a glance

Your video

MP4 / MOV

S3 Upload

Presigned PUT

Whisper

Transcription

Virality AI

Scoring 0–100

FFmpeg

Clip extraction

Your clips

Ready to post

Common questions

How long does processing take?

Most videos under 30 minutes are ready in under 3 minutes. Longer videos or heavy load periods may take up to 10 minutes.

How many clips does one video generate?

Typically 3–7 clips per 30-minute video, depending on content density and how many high-scoring segments the AI finds.

What if my video has no speech?

The AI pipeline relies on speech transcription to find viral moments. Videos without clear spoken content will produce fewer or lower-quality clips.

Can I re-process the same video?

Yes. Each upload consumes tokens, so re-processing deducts tokens again. You might want to do this after buying a higher-tier pack with reframing or subtitles enabled.

Are my videos private?

Yes. Videos and clips are stored in private S3 buckets. Access requires ownership verification on every API request.

Want to dive deeper?See all features·Browse guides
Get Started Today

Start creating viral clips for free

Every account gets 10 free tokens. No credit card needed. Just sign in with Google and upload your first video.