Everyone's taste on the table. One fair pick.

For families and friends with different tastes, ages and diets. Everyone adds three favourite films, shows or artists. An agent asks Qloo's taste graph what each person will like, and picks the dinner spot or movie that leaves nobody behind.

Or start your own huddle

Sunday lunch, three generationsa real run on live Qloo data

A simple average picks

Alimentari

GLMLM

TableFair picks

ima Asa Yoru

GLMLM

Faded bars: Mai is flexible tonight. Qloo sees little difference between these options for this person, so the fair pick protects the others first.

The average leaves Grandma Lan at 50%, her 8th of 15 options. The fair pick lifts her to 68%, and nobody with a clear preference ends up below 65%. Both places are open when the family is going.

How it works

1
Start a huddle and share the link
Dinner spot, movie or TV show. Say when you’re going (we only suggest places open then) and any must-haves like “under $$$”.
2
Everyone adds three favourites
A film, a show, an artist, a book. Qloo’s taste graph connects them to restaurants and titles.
3
Get one pick that leaves nobody behind
The agent scores every option for every person, then picks the one with the best lowest match.

Why Qloo, measured

Random groups on live Qloo data. Against Claude on its own, every pick was graded on a favourite held back from both.

Claude alone, no Qloo: 27%; of its picks break the group's rules: closed, over budget, not vegetarian
Least-matched person, held-back taste: 41 vs 18; percentile with TableFair vs Claude alone; better in 34 of 47 groups
No taste data, most popular pick: 52%; of 120 groups leave someone below 25%; the fair pick, none

Against a simple average the fair pick draws on held-back tastes, and lifts the least-matched person by 14 points on the tastes people gave. Method and the places we don't win are on the Impact page.

Fair, not loudest: Averages hide the person who hates the choice. TableFair lifts the lowest match first and shows what the average would have done.
Shows its working: Every score is computed from Qloo data, and the exact requests are listed. Claude writes the words, never the numbers.
Taste, not questionnaires: Nobody fills in cuisine checkboxes. Loving Spirited Away and Norah Jones is enough to place you.

For teams, platforms and AI assistants

Team dinners, offsites, group trips and shared nights out have the same problem at a bigger scale. Booking, ticketing and travel platforms can call the same fair-pick engine directly: no AI cost per call, a few Qloo lookups, and every result comes with its evidence.

Team-dinner and offsite tools: one venue the whole team can live with
Ticketing and events: shows a group of friends will all enjoy
Group travel: restaurants and activities for mixed-taste groups
AI assistants: add /api/mcp to Claude as a connector and ask for a fair pick, or a link for your friends, right in the chat

Read the API docs

POST /api/v1/fair-pick
{
  "kind": "place",
  "location": "Melbourne",
  "members": [
    { "name": "Ana", "favourites": ["Amélie"] },
    { "name": "Ben", "favourites": ["John Wick"] }
  ]
}

→ { "pick": { "name": "Archie's All Day",
              "meta": "Fitzroy · $$",
              "lowest_match": 0.544, ... },
    "simple_average_pick": { ... },
    "provenance": { "qloo": "live",
                    "calls": [ 7 requests ] } }

Start a huddle

Takes ten seconds. You'll get a link and a QR code to share.