How it works
What happens to your video
Five steps, and you stay in control at the last one.
STEP 1
Transcribe
Speech-to-text with word-level timestamps and speaker pauses.
STEP 2
Detect moments
Find nouns, actions and ideas that benefit from a visual, and sentence boundaries to cut on.
STEP 3
Match footage
Search licensed stock libraries and rank candidates against the surrounding sentence.
STEP 4
Assemble
Place cutaways and overlays, add captions, apply your brand kit.
STEP 5
Review
You review the draft, swap or trim clips, and approve the export.
What the AI can't do
- It can pick a clip that is topically close but not what you had in mind. That is why every clip is swappable.
- It does not understand sarcasm or inside jokes reliably, so literal matches can miss the tone.
- It cannot show your product, your people or your screen unless you upload those clips yourself.
- Transcription can mishear names, jargon and heavy accents. Check the transcript before exporting.
- Stock footage is generic by nature. It will not replace footage that must show something specific.
- It does not guarantee a result that is ready to publish without review.
