Skip to content

Services

AI-training-ready metadata for video archives

Metadata designed for AI training: time-based descriptions of what happens in every clip, so AI teams can select exactly the footage they need. From $0.50 per hour of video.

Timeline 0:03 to 0:07 The dog runs straight at the lens, tongue out, and is met by a reaching gloved hand.

  • 1920x1080
  • 25 fps
  • 12.32 seconds
  • h264
  • 16:9
  • bt709
  • no audio

Why Motion56

Accuracy. Speed. Volume. Trust.

  • Accuracy

    Human checked

    Technical metadata read from the file, descriptive metadata from models, and a sample checked with you before the full run.

  • Speed

    1 working day

    Our reply to every enquiry about an archive.

  • Volume

    79,133 hours

    Of real-world video across our categories: at home, at work, on the road and in the natural world.

  • Trust

    Since 2011

    The year Newsflare, the team behind Motion56, began clearing rights on video. Your archive stays yours.

How it works

Connect. Analyse. Deliver.

  1. We read your archive from cloud storage.
  2. We extract metadata for each asset at your tier.
  3. You receive one structured record per asset.
A young woman sitting on the floor at home, playing an electric guitar
  • guitar
  • musician
  • home
  • wooden floor
  • amplifier
  • sofa

Pricing

Priced per hour of video.

Choose a tier
hours
Do you need vectors?

Vector embeddings let your own systems search the footage by meaning.

Your estimate

$10,000

10,000 hours × $1.00 (Mid)

An estimate, not a quote. Set-up, storage and egress are quoted separately.

Send this estimate

How do you make a video archive AI training ready?

By adding metadata built for AI training selections to every asset. Motion56 reads your archive from cloud storage, generates tags, summaries and transcripts for each file and, at the Complex tier, objects and event timelines, then delivers one structured record per asset. AI teams can then select footage by what happens in it, not by its file names.

Questions

What archive owners ask.

Do we have to license our footage through Motion56?

No. Metadata services stand on their own. If you decide to license later, the metadata is already done.

Can it make an old broadcast or news archive searchable?

Yes. Tags, summaries, transcripts and, at the Complex tier, objects and event timelines are generated for each asset, so the archive can be searched by what is in the footage rather than by the file name.

Does our video need to be in the cloud?

It is quickest if it is. The service reads from cloud storage: AWS S3, Cloudflare R2, Wasabi or Backblaze. If your footage is held anywhere else, we move it onto a server first for an extra fee, and the work then runs as normal.

Is the output reviewed by people?

A sample is checked during scoping and your team can review it before the full run.

Make your archive AI training ready

Book a scoping call