Edge Video AI
Home

AWS Elemental Inference vs Edge Content Enrichment

AWS Elemental Inference brought AI-generated contextual metadata into the AWS encode path. If your stack is already AWS Elemental, that's a strong upgrade. If it isn't, Edge's Autonomous Commerce for ads does the ad-context job as a separate layer that works with any encoder, playout and ad server.

Two different setups

AWS Elemental InferenceEdge Content Enrichment
What it isAWS's AI service alongside Elemental encoding, which also does vertical video, clips and subtitles (AWS product page). Its contextual metadata covers IAB categories, GARM brand-safety signals, detected objects and actions, and shot and scene descriptions.An enrichment layer that reads the live picture and adds context to the ad request
Where it runsMediaLive puts the Inference feed into the SCTE-35 ad markers. MediaTailor fetches the categories at each break.Any encoder, playout and ad server; no re-encode. A light hookup in your ad server or SSAI adds the categories to the ad request.
What reaches the ad decisionIAB Content Taxonomy 3.1 categories and GARM brand-safety signals, fetched by MediaTailor with GetMetadataIAB Content Taxonomy 3.1 category IDs and keywords on the request your ad server already sends
Good fit forTeams already running AWS Elemental MediaLive and MediaTailorTeams on any encoder, playout or ad server

What Edge does, by the numbers

Measured on five live FAST channels, 29 Sep 2026, unless noted.

SSAI stitches. Enrichment decides.

Keep your SSAI and your ad server. Edge adds the context buyers bid on, so the minute on screen can be matched to demand.

Read more

Caught on the hop — how Edge does Amazon-media AI without the scramble

FAQ

What is the difference between AWS Elemental Inference and Edge Content Enrichment?

AWS Elemental Inference is AWS's AI service that runs alongside Elemental encoding. For ads, MediaLive puts the Inference feed into the SCTE-35 ad markers and MediaTailor fetches the categories at each break. Edge Content Enrichment, part of Edge Video AI's Autonomous Commerce for ads, works with any encoder, playout and ad server: it reads the stream you already run and adds IAB Content Taxonomy category IDs and keywords to the ad request your ad server already sends.

Is there an alternative to AWS Elemental Inference for live ad enrichment?

Yes. Edge Video AI's Autonomous Commerce for ads enriches live and FAST ad requests without MediaLive or MediaTailor. It works with any encoder, playout and ad server, with no re-encode. It has been live since 2024.

Do I have to re-encode on AWS to get AI ad context?

Not with Edge. Edge reads five live channels from their public HLS streams, with no re-encode and nothing to change in your encoder or playout. Adding the categories to your ad request is a light hookup in your ad server or SSAI.

What does Edge write on the bid?

IAB Content Taxonomy 3.1 category IDs and keywords from the live picture, on the publisher's ad request as content.cat with cattax. The publisher decides which DSPs see it.

How often does Edge read the screen?

About every six seconds: around 572 reads per hour on average across five live FAST channels (measured 29 Sep 2026), or about 6,700 sellable signals per channel-hour.

What happens if Edge doesn't answer in time?

If Edge is late, the break runs as normal with its usual ads.

Is content enrichment the same as SSAI?

No. SSAI stitches the ad into the stream. Enrichment tells buyers what the minute is about so they can match it. Edge does the enrichment job and keeps your SSAI.

What is Autonomous Commerce for ads?

Edge Video AI's product for placement and purchase: it reads what's on screen, enriches the ad bid, and can sell from what it sees, as on VIDAA V-Shop and MasterChef México Autonomous Recipes.

See how realtime enrichment works →