.png?alt=media&token=92e1d872-53df-40a9-a3df-b00043a3bae4)
AI-powered curation & personalization · streaming & TV
The home screen is the product. Own it.
Personalizer / Curation Hub is the control room for every carousel your audience sees. Content teams compose the experience, choose the AI that fills it, schedule it, test it, publish it — and measure what it returned. No release cycle. No engineering ticket.
The problem
Your recommendation engine is only as good as the surface it lands on.
Most platforms treat the home screen as code. Changing a row means a ticket, a sprint, and a release — so the screen stops moving while the catalogue, campaigns and audience keep moving.
Curation Hub separates the screen from the release. Lanes, algorithms, schedules and experiments become configuration your content team owns directly.
Compose
Build the screen, not the codebase.
Lane groups, lanes, slots and positions — assembled in a visual editor, rendered identically by your apps.
Create, position, size and title every carousel, then serve the result to every app through one render call.
Lane groups as screens
Compose home, kids, sports or any surface as an ordered set of lanes.
Fixed positioning
Pin commercial commitments while letting the rest of the screen personalize.
Localized titles
Resolve the right title for each locale at render time.
Personalize · the AI engine
Pick the algorithm the way you pick a playlist.
Every lane points at a reusable personalizer configuration.
Choose the algorithm, catalogue slice and audience. Change the recipe and the intelligence behind the row updates instantly.
Algorithm per lane
Recommendations, similarity, trending and editorial picks share one publishing path.
Catalogue filters
Constrain what each algorithm may surface using your own metadata.
Audience segmentation
Target each configuration using consumption-based audience profiles.
Schedule & experiment
Run the screen like a schedule, prove it like an experiment.
Time-boxed slots, native A/B variants and guaranteed fill — all declarative.
Launch premieres and campaigns automatically, test a hypothesis on a share of the audience and ensure no carousel is ever empty.
Slot schedules
Define start and end times for any individual slot.
Native A/B variants
Attach a variant lane and the audience share that should see it.
Fallback lanes
Keep every row full when an algorithm cannot reach its threshold.
Govern & publish
Nothing reaches your audience until someone decides it should.
Draft and published are different states. Preview is the real thing. Every change is attributed.
Production keeps serving the last published version while your team previews pending changes through the real personalization pipeline.
Explicit publishing
Saving never changes production; publishing is a deliberate action.
True preview
See output for a real profile, locale and moment before going live.
Enterprise identity
Role-based access, SSO and complete change attribution.
Measure
Curate. Publish. Find out.
Lane-level performance and consumption analytics in the tool that made the decision.
Close the loop with clicks, conversions, watch time, completion and audience performance by lane and screen.
Lane performance
Clicks and conversions by lane, screen and day.
Consumption metrics
Playbacks, viewers, hours watched, devices and completion.
Audience analysis
Rank performance by package, period, status and region.
For engineering teams
Two APIs. One contract. Every brand.
Your apps talk to stable configuration and runtime APIs. Every brand runs its own isolated instance with the same contract and interactive documentation.
GET /v1/curation-hub/capabilities
POST /v1/curation-hub/jobs
GET /v1/curation-hub/jobs/{id}
200 · application/jsonProduct spec
Personalizer / Curation Hub
Make every screen move at the speed of your audience.
See Personalizer / Curation Hub working with your own catalogue, audience and data.
