Book a demo
Menu

Metadat-AI for Iconik

iconik

Metadat-AI reads your video, images and audio and writes what it finds back to your Iconik assets as time-coded metadata: scenes, on-screen text, transcripts, faces, logos, offensive content, industry ratings and sports actions. Your team stops logging footage by hand and finds the moment in seconds.

Why connect Metadat-AI

Media libraries grow faster than any team can log them. Someone needs the clip of the keeper saving the penalty, but nobody logged it, so a producer ends up scrubbing through match footage. Metadat-AI closes that gap by reading the media itself.

  • Find the right moment in plain language Ask for "the scene where the keeper saves the penalty" and find the matching scene, the line in the transcript, and related images and sound clips (documents coming soon). Results come with 87% less metadata noise than tagging one frame per second.
  • Your media stays in your environment Run Metadat-AI in your own cloud account or on your own hardware, and your files are never sent to us. A hosted option is also available.
  • Up to 50% lower total cost Simple per-hour pricing with no egress fees. With commitment and prepayment discounts, that comes to up to half the cost of comparable solutions.
  • AI you can configure Choose which models run on which content, such as a football template for match footage and a drama template for scripted content, and add compliance screening against UK broadcast and advertising rules.
  • Hands-on help Our team configures and tunes models and workflows with you, and builds custom models for your content.

Who it is for

Any Iconik team that needs fast, automated metadata, for example:

  • Sports teams and rights-holders Players, sports actions and sponsor logos tagged as time-coded metadata, ready for sponsor reporting
  • Broadcasters and large archives Scene descriptions, on-screen text, transcripts, content ratings and offensive-content flags, so the library is searchable by what happens in it
  • Post-production and creative teams Rushes tagged as they are ingested, so editors find the shot by describing it
  • Brands, agencies and advertisers Risks such as HFSS food, alcohol and gambling flagged against UK advertising rules, with the exact moment
  • Compliance and legal teams Offensive content, age-rating risks and AI-generated video flagged by timecode
  • Quality and inspection teams Rust, cracks and defective parts flagged in drone and production line footage, graded by severity

How it works

Illustrative example. Metadat-AI receives only a short-lived link. No media.

Your environment (your cloud or your hardware) short-lived link only instructions iconikyour media library Your storagemedia stays here Search Your databaseoptional Analysisyour cloud or your hardware Metadat-AIa short-lived link only.No media. Scene Logo Speech Your environment (your cloud or your hardware) short-lived link only instructions Metadat-AIa short-lived link only.No media. iconikyour media library Search Analysisyour cloud or your hardware Scene Logo Speech Your storagemedia stays here Your databaseoptional

An Iconik admin creates an application token and a Custom Action that puts Metadat-AI in the Iconik right-click menu, or a webhook that runs it automatically on new assets. Your team sends assets or collections to Metadat-AI. Metadat-AI reads each asset's Iconik proxy once through a short-lived secure link, and writes the results back to the asset as time-coded metadata, ready to search in Iconik.

Setting it up

Our team, or an approved reseller, usually does the setup with you.

  1. Talk to us. We agree your use cases, the models to run, and where processing happens.
  2. Create the token. In Iconik, create an App ID and Auth Token for Metadat-AI and share it securely with our team.
  3. Create the Custom Action or webhook. In Iconik, point it to your Metadat-AI endpoint for the job template you want to run. We provide the URL and settings.
  4. Map the metadata. We map Metadat-AI's output to your Iconik metadata view.
  5. Ready to go. Start running jobs from Iconik.

Customer story: a leading European football club

Tagging every player in every clip was a full-time manual job for the club's media team. Off-the-shelf models struggled to identify players from extreme angles and from behind, and EU rules limit biometric analysis.

  • Players identified automatically from any angle Using face recognition where permitted, and shirt numbers and printed names everywhere else, with no manual logging
  • Ready for sponsor reporting Results filtered to the club's own players, with sponsor logos tagged in every clip
  • Within the rules Built to support EU GDPR and the EU AI Act, so the club can tag every player at scale

Security and data

  • Metadat-AI reads the Iconik proxies, once each, through a short-lived secure link from your cloud storage.
  • Processing can run entirely in your own cloud account or on your own hardware.
  • Your content is never used to train our models.
  • Metadata is stored in Iconik, not held by Metadat-AI as a system of record. The option to keep search data in a database you control is planned.
  • Player recognition from shirt numbers and names processes no biometric data. Face recognition is optional and off by default.

FAQ

What do we need in Iconik?

Admin access to create a Custom Action or webhook, an application token for Metadat-AI, and a metadata view with the fields you want filled. We can help set this up.

Does Metadat-AI read our original files?

No. It reads the Iconik proxies.

Can we choose which models run?

Yes. Job templates decide which models run on which content.

Where does processing run?

In your own cloud account, on your own hardware, or in our hosted environment on AWS in Ireland. For enterprise customers who host everything themselves, Incognito mode means none of your media or metadata comes to us.

Can developers build on the API?

Yes. Swagger API documentation and an API Playground are available inside the platform, and our Claude Code skills help developers build integrations with our APIs.

About integrations