नेपाली भिडियोका लागि
One long video. Ten clips that travel.
Drop in a Nepali video and get vertical clips back - each one scored, reframed, and captioned in Devanagari that actually composes.
No card. Thirty minutes of source video on the free plan.
नेपालको सबैभन्दा अग्लो
00:38यो कुरा कसैले भनेन
00:52काठमाडौंको समस्या
01:04मैले सिकेको पाठ
00:47Three things have to go right.
It listens in Nepali
Transcription runs on the actual audio, with your own glossary of names and places so proper nouns come back spelled the way you spell them.
It picks the moments
Every candidate span gets scored on hook, flow, value and trend - and tells you why, so you can disagree with it.
It frames and captions
Face-aware vertical crop, then captions burnt in with conjuncts composing correctly. That last part is where most tools fall over.
The actual problem
Devanagari is not a font problem.
Nepali stacks consonants into conjuncts and hangs vowel marks above and below the line. A renderer that treats text as a row of glyphs breaks them apart - you get the right letters in the wrong shapes, which reads as broken to anyone who speaks the language.
Getting this right end to end - in the editor, and burnt into the exported frame - is the reason this exists rather than a wrapper around someone else's clipper.
Rendered output
सगरमाथा संसारको सबैभन्दा अग्लो हिमाल हो