AI Recommendations that feel personal

Daily Match% picks tuned to your group, mood, time and platforms. The entire ranking is driven by a highly sophisticated AI.

92% MatchBecause: feel-good, <2hOn Netflixkey 'pills.with (en)' returned an object instead of string.
Keekup — Recommendations preview

Why our recommendations are smarter

👥
Who you're watching with

Different picks for date night, family or solo — the AI models co-viewing preferences, not just individual taste.

🎛️
Context & constraints

Mood, runtime window, content rating, languages/dub/sub, even your current device. The engine scores what fits now.

📺
Streaming availability

Only show titles available on your providers/regions to avoid dead ends.

🧠
Personal taste graph

Learns from your adds, ratings and Top 5 to map fine-grained affinities across genres, tones and pacing.

🌍
Multilingual & locales

Understands original language, dubbed tracks and subtitles — great for mixed-language groups.

📈
Learns in real time

Every swipe/save/refusal updates future picks. Balance between safe bets and fresh discoveries.

How it works

1
Set the session

Pick who’s watching, mood and constraints (time window, age rating, language…).

2
The AI composes a pool

It builds a candidate list that matches your availability and context.

3
Ranked by Match%

Each title gets a Match% score based on fit and your taste graph.

4
You interact — it adapts

Swipe, save or rate; the next picks instantly reflect your choices.

Daily Match% picks

Fresh suggestions every day. Quick preview, instant reasons, and a tap to play — no endless scrolling.

Transparent & private

Clear reasons

We surface ‘because’ chips (mood, runtime, provider, languages…) so you see why a title fits.

Privacy by design

Your data stays minimal and is used only to improve your recommendations. No selling, no surprise sharing.

Recommendations FAQ

What is Match%?

A score estimating how well a title fits your current session (who’s watching, mood, constraints) plus your historical taste.

Do I need to rate a lot to get good picks?

No. You’ll get solid results from context alone; ratings and saves just make them sharper, faster.

Does it work for groups?

Yes. The engine blends preferences and constraints from multiple people to maximize shared enjoyment.