Everything you need to turn a long stream into clips that work.
From fetching the VOD to publishing on TikTok, Nysos automates the whole vertical-clip production chain. Here's everything the tool can do.
Multi-signal highlight detection
Cross-references audio peaks, chat spikes, emote intensity and transcript topic shifts to find a stream's real best moments. Not a blind cut every 60s.
4-axis A→F scoring
Every clip is graded by an LLM on Hook, Flow, Value and Trend. You see why the model picked the moment, not a black box.
Learn moreAutomatic karaoke captions
Deepgram nova-3 transcription with word-by-word timestamps. Per-word karaoke animation, viral-ready fonts (Bebas Neue, Anton, Impact), keyword highlighting.
9:16 format for TikTok / Shorts / Reels
Automatic vertical reframing with face detection. Podcast mode (2-speaker split-screen), entertainment mode (1 face), blurred-background mode.
Twitch & Kick Live Watcher
Monitors one or more channels in real time. As soon as a stream ends, Nysos automatically launches clip generation. Set and forget.
Learn moreFormat Converter
Import any horizontal video (16:9, 4:3, raw) and convert it to publish-ready 9:16. Essential for recycling your existing catalog.
Caption Remover
Removes burned-in captions (open-captions) from a video you re-upload or recycle. No more double subtitles.
Timestamped transcription
Transcribes the source with precise timestamps to power scene analysis, captions and transcript exports — without promising perfect speaker attribution.
Direct TikTok / YouTube Shorts publishing
Connect your accounts once, then publish clips in one click. Title, description, privacy, scheduling — without leaving the dashboard.
Built-in YouTube analytics
Once published, Nysos automatically pulls YouTube metrics (views, watch time, retention) to compare what actually works for you.
Transcript download
Download your source's full transcript as .txt (readable) or .json (with word-by-word timestamps). For editorial reuse, custom captions, or archiving.
Fast single-pass pipeline
One LLM analysis in a single call over the whole transcript + signals: faster, cheaper, better global context. A 4-hour stream processes in 8-12 min.
Ready to try?
60 free source minutes at signup. No card required. Generate your first clip in under 5 minutes.