Seatback

Judge mode

The question: of the titles you can license, which suit the passengers on a given route? Open an answer below.

Example routes. One click opens an answer: replayed from a stored run when this deployment has one (no new Qloo calls), otherwise computed from synthetic fixtures.

Build your own route slate

From To

Passenger mix 50% origin taste / 50% destination taste Licensable catalogue one title per line (tv:, podcast:, music: prefixes) or CSV rows: type,title,year

Import CSV

Rules Family-safe screens only (US ratings) Suggest films released since Programming theme (optional, e.g. holiday, sports)

Films TV Podcasts Music

Custom runs are rate-limited per visitor. With no Qloo key they run on synthetic fixtures and your input is not sent anywhere. With a key, Qloo queries built from your input go to Qloo and its responses are cached privately for up to 14 days; a language model, if configured, also sees your route and titles. Your catalogue text is not saved.

Program a route, not a fleet.

The question it answers: of the titles you can license this cycle, which suit the passengers on this route? Seatback reads both end cities from Qloo's taste data and ranks your catalogue for the route, before you license.

One click opens an example answer: replayed from a stored run when this deployment has one, otherwise computed from synthetic fixtures (the trace says which). Nothing to type.

Honest limits, up front

Affinities are Qloo's relative taste signals for a place. They are not viewing forecasts, and nothing here has been validated against airline viewing data.
Where the data mode says MOCK, every score comes from a synthetic fixture and every title is invented. It shows how the tool works, not what Qloo says.
Matches the run is unsure about are listed with candidates and confidence; it does not quietly pick one.
Some endpoints (trending, demographics) may not exist on the hackathon host. If so, that part is reported as unavailable, not invented.

Airlines license entertainment by the cycle and program it by route, yet in-flight viewing data is patchy across fleets and arrives after the licences are bought. Seatback reads both ends of a route from Qloo's taste graph before you license, ranks the catalogue you can actually buy, and shows where else on your network the slate fits.

How it works, in four steps

Read each city. Qloo's open location read shows what each end favours, and names its own top local artists and shows.
Score your catalogue twice. Those local anchors score your titles in two independent readings. Titles the readings disagree on are withheld, not guessed.
Balance the slate. Blend both ends by your passenger mix, keep variety, respect family-safe rules, and show the Qloo call behind every number.
Extend it to your network. Qloo's cross-domain graph turns the slate's audience into the destinations it favours, checked twice, so you see which other routes the same licences suit.

Why this matters (sourced): one provider's 2018 content revenue, its airline customers, and an executive on data demand

$315Ma year in airline content licensing and services at a single provider (Global Eagle, now Anuvu), 2018, one provider (10-K)

150+airline customers served by that one provider today (Platinum Equity, Anuvu profile)

"Demand for data"is what Anuvu's SVP of media & content reports; Anuvu has partnered with Parrot Analytics for content-demand insight (Runway Girl Network, 2024)

Qloo taste data has reached IFE before: West Entertainment's Jetset AI (2020), built on Qloo's platform to personalise airline content. Seatback is the programmer's side: route by route, before licensing, on your own catalogue. Sources are listed in the README.

Judge mode opens an example answer. With no Qloo key it is computed from synthetic fixtures when you click; it is a replay of a stored run only when the trace says "Replayed from a stored run" (and then the pill says LIVE if it was recorded from live Qloo data).