For business

The Taste API: recommendations for streaming, books and games

Boxlist matches films, TV, books and games to each person’s taste. The same recommendation engine can run inside your product, on your catalogue: a short taste test, a match on every title, Similar in feel and picks for two.

Talk to us See an example

Early access. We’re shaping the API with a small number of pilot partners.

Built on a live recommendation engine

Everything below already runs on Boxlist for consumers. The API is the same engine.

Sample taste type

Hard Hitter

SlowFast
Tried & testedFresh / original
UnsavouryWholesome
Light / accessibleWeighty / heavy
Imaginary worldTrue to life

+ 5 more dimensions

Taste Fingerprint™
A taste profile on ten characteristics, from pace to emotional punch, plus four Taste Types. A two-minute test builds it.
A match on every title
Five plain bands, Perfect match to Not for you, relative to each person, with the reasons shown.
Similar in feel
Titles that feel alike across the ten characteristics, not just the same genre or cast.
Joint taste
How alike are two people, and what suits them both. Each side sees bands, never the other’s answers.
Moods
Feel-good, Thrilling and smart, Mindbender: moods built from the same characteristics.
Films, TV, books and games
One fingerprint matches across all four. Where to watch for the UK and US.

Recommendation and personalisation APIs

Take one piece or several. Each works on your own catalogue.

Taste Match API

Score your titles for one user.

Send a fingerprint (or a few liked titles) and title ids; get back a band per title, best first.

Similar in feel API

Who it’s for

Streaming services

Onboarding: a short taste test instead of “pick three genres”
Cold start recommendations: good rows from the first session
Personal rows and moods, with a reason on each title

Publishers and bookshops

Book recommendations matched to a reader’s film and TV taste
“If you liked this series” across media
Taste-led newsletters and landing pages

Game stores

Game recommendations that match what a player watches, not just plays
Similar in feel across genres
Help players skip buys that don’t suit them

Telcos and aggregators

One taste profile across many apps and services
Watch-together picks for households
A shared front door for bundled catalogues

How it works

From cold start to ranked rows, on your own catalogue.

Learn the taste
Your users take the taste test in your app, or you send a few titles they liked. We return a Taste Fingerprint™.
Score your catalogue
We score your titles on the ten characteristics from your own metadata, once, and keep them up to date as titles are added.
Match and rank
Call the API for a band per title, Similar in feel or a pick for two, and use it in your rows, search and emails.

An example API call

Score three titles for one user. Illustrative: the shape of the API, not a live endpoint.

Request · POST /v1/taste-match

{
  "user": "u_8f2c",
  "fingerprint": {
    "pace": 4.6,
    "originality": 4.1,
    "wholesomeness": 2.2,
    "weightiness": 3.9,
    "realism": 3.1,
    "intricacy": 4.4,
    "linearity": 2.6,
    "styleAndVisuals": 2.4,
    "scale": 3.8,
    "emotionalPunch": 4.2
  },
  "titles": [
    "tt0816692",
    "tt1375666",
    "tt0110912"
  ]
}

Response

{
  "user": "u_8f2c",
  "tasteType": "Thrill Seeker",
  "matches": [
    {
      "title": "tt1375666",
      "band": 1,
      "label": "Perfect match",
      "reasons": [
        "Smart / intricate",
        "Fast"
      ]
    },
    {
      "title": "tt0816692",
      "band": 2,
      "label": "Strong match",
      "reasons": [
        "Emotive / stirring",
        "Big / varied"
      ]
    },
    {
      "title": "tt0110912",
      "band": 3,
      "label": "Mixed",
      "reasons": [
        "Surreal / non-linear"
      ]
    }
  ]
}

PerfectStrongMixedWeakNot for you

Bands are relative to each person, so a Perfect match is among their best fits.

Privacy-first personalisation

We never sell personal data. Your users’ taste profiles stay yours.
UK and EU GDPR: you are the controller and Boxlist acts as your processor, under a data processing agreement.
Fingerprints need no name or email: a pseudonymous id is enough.
Joint taste shares bands only. Neither person sees the other’s answers or scores.
Audience insights are aggregated and anonymous.

Questions from product teams

How is this different from a collaborative filtering recommendation engine?

Collaborative filtering needs lots of viewing history and learns “people who watched X also watched Y”. The Taste API matches on how a title feels, across ten characteristics, so it explains each pick and works on niche and new titles with no history yet.

How does it handle cold start?

A new user takes a two-minute taste test, or you send a few titles they liked, and gets a Taste Fingerprint™ straight away. New titles are scored from their metadata, so they can be recommended from day one.

Do you need our user data or viewing history?

No. A pseudonymous id and a fingerprint, or a few titles the user liked, are enough. No name or email is needed, and you stay the controller of your users’ data, with Boxlist as your processor.

Can it work on our own catalogue and metadata?

Yes. We score your titles from your own licensed metadata and keep the scores up to date as titles are added. Films, TV, books and games can share one taste profile, so a reader’s or player’s picks can draw on what they watch.

How do we integrate it?

Through a JSON API that returns a match band per title, Similar in feel lists and picks for two, plus an embeddable taste test widget for onboarding. You decide where the results appear: rows, search, emails or landing pages.

Is the API available now?

It is in early access. We’re shaping it with a small number of pilot partners, so tell us about your catalogue and users and we’ll reply with what a pilot could look like.

Talk to us about a pilot

Tell us about your catalogue and your users. We’ll reply with what a pilot could look like.

Email team@boxlist.com