Arena Pro

Find any moment with natural-language search

People remember a guest, a phrase, or a graphic — not a filename. Arena Pro searches the understanding of the file.

Arena Pro natural-language search resolving to an exact moment inside an asset

The problem

Filename search assumes the person who stored the file is the person who must find it. In a video archive that is rarely true. The request is usually a person, a phrase, or a graphic.

How Arena Pro solves it

Natural-language and semantic search run across assets. AI-generated metadata is attached on ingest. Results can point at a file or at a moment.

Arena Pro video OCR with recognized on-screen text on the timeline
  1. 00:12:04

    Face

    Juan Massana

    Recognized face, 15s appearance

  2. 00:11:58

    OCR

    BANCO Mediolanum

    On-screen text, frame-level timecode

  3. 00:12:11

    Transcript

    “…the archive that's always ready”

    Synchronized speech

  1. Ingest the file or live chunk.
  2. Wait for analysis — or keep working and use the in-app notification when it finishes.
  3. Ask for the person, phrase, or on-screen text you remember.
  4. Open the inspector or Studio at the timecode.
  5. Clip, share, or publish from that point.

AI Search

Search that understands the content

AI Search

Use natural language to find any content in your library.

Manual Filtering →

Ask

Describe what you need in plain language. Transcriptions, faces and recognized text are all searchable — not just filenames.

  • Ask for what you remember

    Natural-language and semantic search return the asset or the moment, not a folder path.

  • Every appearance of a person

    Face recognition marks who is on screen so you do not scrub an hour to confirm a guest.

  • Text that appeared on screen

    Video OCR stores on-screen words with frame-level timecodes — lower thirds, scores, signs.

  • From long-form to suggested clips

    AI-assisted smart clip suggestions give editors a starting cut, not a blank timeline.

  • Landscape to vertical

    Subject-aware crop produces 1080×1920 deliverables inside Studio.

  • Images without a side tool

    Background removal, auto-description, and keyword suggestions stay on the asset.

How it works

Analysis runs as jobs on ingest or on demand. Results attach to the asset as metadata and timeline tracks. People remain in control: permissions, review states, and human edits sit on top of machine output. Vendor and model names are not published here unless product marketing approves them.

Bring one workflow. We will open it in the product.

Talk To Us