[VORTEX Portal] AI Hub

Overview of AI Hub

AI Hub is a centralized platform within VORTEX designed to unify and streamline all AI-powered surveillance features. It provides users with a single interface to access Think Search, Deep Search, Event insight, and Case Vault, making surveillance operations more efficient and intuitive.

Note: AI Hub will be released initially as a beta version

Think search

Think search allows users to quickly find relevant video footage based on specific objects or contextual information. It includes two main modes:

• Powered by VLM (Vision-Language Model), this feature enables users to input simple, natural language queries (e.g., "man with red shirt" or "white SUV at entrance") to locate matching events or best-shot images, making search more flexible and intuitive.
• Best practice
  • We suggest structuring queries as follows:
    • [Person or vehicle] + [action/descriptor] + [optional connector] + [optional second descriptor] + [optional location]
      • Example:
        • Person wearing a black jacket with sunglasses at 7 am on the street.
        • Person carrying a large backpack with headphones in the library.
  • Supported languages
    • English (en), Simplified Chinese (zh), Spanish (es), Hindi (hi), Arabic (ar), Bengali (bn), Portuguese (pt), Russian (ru), Japanese (ja), Punjabi (pa), Malay (ms), Tamil (ta), French (fr), German (de), Turkish (tr), Korean (ko), Urdu (ur), Hausa (ha), Greek (el), Dutch (nl), Hebrew (he), Italian (it), Indonesian (id), Serbian (sr), Swahili (sw), Marathi (mr), Belarusian (be), Slovak (sk), Thai (th), Finnish (fi), Ukrainian (uk), Swedish (sv), Malayalam (ml), Punjabi (pa), Czech (cs), Bulgarian (bg), Irish (ga), Georgian (ka), Lithuanian (lt), Amharic (am), Albanian (sq), Armenian (hy), Azerbaijani (az), Catalan (ca), Macedonian (mk), Maltese (mt), Malayalam (ml), Slovenian (sl), Serbian (sr), Kazakh (kk), Mongolian (mn), Latvian (lv), Estonian (et), Lithuanian (lt), Georgian (ka), Kikuyu (ki), Sorani Kurdish (ckb), Hakka Chinese (hak), Ossetian (os), Cherokee (chr), Kurdish (ku), Greek (el), Polish (pl), Latin (la), Greek (el), Sinhalese (si), Hebrew (iw), Belarusian (be), Hakka Chinese (hak), Spanish (es), Chinese (ch), Spanish (es).
    • Known Limitations:
      • Accuracy decreases when searching for multiple people simultaneously.
      • Difficult to identify small objects (e.g., phones, wallets, earbuds).
      • May return imprecise results for action-based scenes (e.g., running or waving).

Deep search

• This traditional filter-based search allows users to refine results by object type (e.g., person, vehicle), appearance attributes, tags, time range, and camera location. It is useful for precise filtering when specific conditions are known.

Event Insight

Event Insight provides users with a visual summary and detailed breakdown of key events detected by AI. It helps identify the person or object responsible for specific incidents, understand behavior patterns, and spot anomalies.

• Known Limitations:
  • Event source from NVR does not support the "Add to case" feature

  • When data is not ready, a tip appear to notify you that the data is not ready for further operation.

    

  • The difference in data information between VORTEX camera and VORTEX Connect
    • No best shot displayed on the event from VORTEX Connect.
    • No detection area/line displayed on the event from VORTEX Connect.
    • No path displayed on the event from VORTEX Connect.
    • Only archive and send feedback are supported for events from VORTEX Connect.

Case Vault

Case Vault is designed for securely storing and organizing critical footage related to investigations. Users can group video clips, add notes, and collaborate with team members — all within a dedicated case management space. This improves workflow for incident review, evidence sharing, and long-term retention.

Detailed information about the Case Vault, please refer to this article.

Language moderation