The AI engine suite built for broadcast

AI automatically segments video into frames, shots, and scenes, organizing each segment into structured metadata. It runs on-premises, so your content never leaves the facility.

• 01Object detection
  classifies and auto-tags more than 116 classes — people, vehicles, animals, backgrounds — using panoptic segmentation.
• 02Video summaries & scene descriptions
  generative AI writes a human-readable description for every scene.
• 03Natural-language search
  unifies face, object, STT, and scene data so you can search the archive in everyday language.

Every word spoken on air becomes searchable text — automatically, accurately, and in real time.

• 01Multi-engine STT hub
  choose from Google, Amazon Transcribe, Naver Clova, OpenAI Whisper, and Daglo.
• 02Speaker diarization
  identifies who said what, and when, with timecode.
• 03Keyword search within captions jumps straight to edit points, with AI caption editing, automatic summaries, and caption file downloads.
• 04Deploy as cloud SaaS (usage-based billing) or fully on-premises with Whisper.

Find every appearance of an on-air talent across the entire archive in seconds, not hours.

• 01Automatic face extraction
  detects faces via landmark analysis and clusters the same person automatically.
• 02Image-based search
  upload a single photo and instantly find every matching appearance.
• 03Person timelines
  shows each person's appearances on a shot-level visual timeline.

Automatically detects, indexes, and searches lower thirds, channel chyrons, and on-screen graphic text.

• 01Full-frame analysis mode
  detects every text region on screen automatically, with no setup.
• 02Region-select mode
  drag to define regions of interest and extract text from exactly the areas you want.
• 03Supports multiple languages
  Korean, English, Chinese, Japanese, and more — across a wide range of fonts.

A prompter that listens to the presenter, matches their speech to the script in real time, and scrolls automatically.

• 01Real-time voice matching
  analyzes the presenter's speech and tracks the script position automatically.
• 02No operator required
  scrolls fully automatically, with no dedicated staff.
• 03Deploys on-premises or in the cloud (including Whisper Live).