Moûsa

An entry to the Qloo Agentic Hackathon, 2026

Design for an audience from what it actually loves.

Ask an agent who an audience is, and it answers from its own idea.

Moûsa is a plugin for the design agents teams already use: Claude Code, Codex and Kiro.

Open the live site Read the source

Transcript, the problem

A plausible design, from an untested idea.

Q.

What happens when an AI agent is asked to design for an audience?

A.

It designs from its own idea of that audience. That idea is never tested, so the design looks plausible and generic.

Q.

How close does that idea get?

A.

In 11 recorded Moûsa sessions in Claude Code, Codex and Kiro, the agents' sealed guesses named 22 of the 714 works Qloo measured for their audiences. 692 of them (96.9%) were missing from the guesses.

Q.

Is that one agent having a bad day?

A.

In a pre-registered benchmark of 20 audiences, the sealed guesses of three agents named 6.8% (Claude Code), 6.6% (Kiro) and 4.7% (Codex) of the works Qloo measured.

The blind spot A field of 714 small dots in 21 rows of 34. Only 22 of them are filled: the measured works that the guesses named.
a work Qloo measured in 11 sessions   one the guesses named. Positions are illustrative.
Missing from the guesses
692 of 714
Share of measured works missed
96.9%
Named by the guesses
22 of 714

How it works, in five steps

Guess first. Then measure. Then design from the difference.

  1. 1

    Plan.

    FRAME 01PLAN

    Moûsa picks the taste questions for the medium: which Qloo domains to measure (films, music, books, TV, brands, places, games) and which tags.

  2. 2

    Freeze the guess.

    FRAME 02SEAL

    Before any Qloo data, the agent writes its own prediction of what the audience loves, and the design it would make from that alone. The prediction is frozen and sealed with its SHA-256.

  3. 3

    Measure.

    FRAME 03MEASURE

    Moûsa asks Qloo what the audience loves in each domain, which style and aesthetic tags it carries, and which qualities its measured works share. A session sends at most 24 requests, and the answers are cached for a week.

  4. 4

    Design from the difference.

    FRAME 04DESIGN

    The agent writes a Design DNA: concrete devices for palette, type, motion, illustration and voice, each anchored to what Qloo measured. The surprises must move it away from the frozen sketch. Moûsa checks every device in the project's own design system (CSS custom properties, token files or a Tailwind theme), and the agent reviews screenshots of what it built.

  5. 5

    Report.

    FRAME 05REPORT

    The Taste Provenance report sets the guess beside the measurement, marks each measured name as new or predicted, and follows it to each device and to the screenshots reviewed.

The evidence, and what each kind does not show

Five findings, each with its limit behind it.

Turn each card to see what it does not show.

  • 01 · The blind spot

    692 of 714 were not in the guesses.

    In 11 sessions, 692 of 714 measured works were not in the guesses, and 22 of the 245 works the agents guessed were measured.

    Backstage

    It does not show that the design is better.

  • 02 · The benchmark

    Twenty audiences and three agents.

    Claude Code, Codex and Kiro. The protocol and all 60 sealed predictions were committed to the public repository before Qloo measured any audience. The guesses named 6.8% (Claude Code), 6.6% (Kiro) and 4.7% (Codex) of the works Qloo measured, and 14% to 16% of the works they guessed were among those measured. Both pre-registered hypotheses held for every agent. The agents knew music and TV best (about 10%) and films, brands and books least (2% to 6%).

    Backstage

    It measures what the agents assume, not whether a design is better.

  • 03 · The mechanism

    One frozen guess, built twice.

    Four recorded pairs were each built twice from one frozen guess, once with Qloo's measurement and once from the guess alone. The distances between the two designs were 0.53, 0.57, 0.58 and 0.60, on a heuristic scale from 0 (alike) to 1 (unrelated). Every device of each Qloo design rests on a name the guess missed.

    Backstage

    That scale is not a quality score.

  • 04 · Beyond the model's own variance

    Alike when blind. Apart after Qloo.

    In repeated pairs of one website brief (12 generations), the blind sketches of both arms were alike (p = 0.81). After Qloo, the typography, structure, palette and overall distance separated (exact permutation test, p = 0.005 each). For game characters, the finished characters separated (p = 0.0079) and the blind sketches did not.

    Backstage

    It shows difference, not preference.

  • 05 · The blind vote

    Which one would the audience rather have?

    Visitors see both designs of a pair without knowing which used Qloo, and choose the one the audience would rather have.

    Backstage

    Voters choose themselves, and there are few.

Three agents, one engine

Use it where your team already works.

Claude Code

  1. /plugin marketplace add tritonsan/mousa
  2. /plugin install mousa@mousa

Comparisons run as two subagents at once.

Codex

  1. codex plugin marketplace add tritonsan/mousa
  2. codex plugin add mousa@mousa

Codex's own image tool can paint the pictures a plan asks for.

Kiro

  1. kiro-cli powers install plugins/kiro

From a clone of the repository.

Each person uses their own Qloo API key, entered on a local page, never in the chat.

Try it

Name an audience. Watch the guess meet the measurement.

The live try

Runs Moûsa in Kiro on the hosted server for an audience you name.

  • Step one, under a minute, shows Kiro's sealed guess beside Qloo's measurement.
  • Step two, about five to ten minutes, builds both designs.
  • The access code is in the submission's testing instructions.

d1n6iqgwqs2043.cloudfront.net

The gallery

Recorded sessions from all three agents. Each shows the design, its record and how to run it again.

The blind vote

Asks which design the audience would rather have.

How it is built

Small, local and tested.

  • One MCP server in Node 22, with no dependencies, packaged for Claude Code, Codex and Kiro from one source.
  • Four skills (taste-brief, design, compare, connect), 13 tools and 33 open-licence font families.
  • A write gate holds a design's files until its Design DNA is valid, in Claude Code and Codex.
  • Captures in a headless Chromium that can reach only the page it captures.
  • More than 500 tests, with no network or model call, run on Ubuntu and Windows.
  • The hosted site runs on one AWS CloudFormation stack: an instance behind CloudFront, with the keys in Secrets Manager.

Qloo's data

Qloo's answers stay on the machine that asked for them.

The site and the public repository show the names Qloo measured, never Qloo's ids, scores or descriptions.

This page

Made twice, from one sealed guess.

This page was designed by Moûsa in Claude Code, twice, from one sealed guess: once with what Qloo measured about its audience, and once from the guess alone.

Its audience is described to Qloo as people who love This Week in Startups, The Original Kings of Comedy, the Los Angeles Dodgers, Chelsea F.C. and Live Nation: works and names the hackathon's judges are known for. That description tells Qloo who the audience is. It is not a claim about any judge's own taste, and no judge endorses this entry.