I
Evidence or abstention
Every factual claim cites the Qloo call that produced it. When evidence is insufficient, the system says so and refuses to force a recommendation.

Cultural intelligence for independent cinema
Investigate a film’s cultural relationships, examine the evidence, and develop a positioning hypothesis worth testing.






A title, logline, genre, the references that shaped it, and what you want the release to do. Nothing else is assumed.
Resolved references, a taste profile, adjacent worlds ranked — every step recorded with a provider call id.
Findings are typed FACT, INTERPRETATION, HYPOTHESIS or UNKNOWN. Facts cite the call that produced them.
A positioning proposal you edit and confirm. What was proposed and what you changed stay separately on the record.
Record what actually happened — including zero, partial or missing. A missing measurement is never a negative result.
Supported, contradicted or inconclusive against the criteria you pre-registered. Correlation is never causation.

Why this is not another analytics dashboard
I
Every factual claim cites the Qloo call that produced it. When evidence is insufficient, the system says so and refuses to force a recommendation.
II
Cultural affinity measures similarity between expressed interests. It is not proof anyone will watch, enjoy or share your film.
III
Hypothesis → experiment → observation → evaluation, with append-only history. You can always trace what was decided, on what evidence, and what the outcome showed.
Photo · Uday Veeru / Pexels · atmospheric reference, not a film still
Reading the evidence
The agent decides with a fixed, auditable plan — resolve, analyse, expand, rank — and stops when the evidence runs out. The model may propose. The evidence constrains it. You decide.
Returned by a recorded Qloo call — cites its call id.
Reasoning over facts — labelled, never sold as source data.
A positioning decision to test, not an established truth.
Openly unresolved — never filled with plausible-sounding content.
Product signature
At the centre: the film, as type. Around it: the cohorts the analysis actually returned. Select a node to inspect claim type, source, and limits. Empty is empty. Fixture is labelled fixture.
Inspect a labelled fixtureCultural Multiverse
Anchor, Bridge and Frontier as visual strategies. This section demonstrates the interaction design only — the evidence-backed arena runs inside Creator Studio against a completed investigation.
CONCEPT PREVIEW — ILLUSTRATIVE DATA — NOT A LIVE QLOO ANALYSIS
Anchor
The composition is stable, centred and editorial. It privileges resolved comparables and refuses to inflate a weaker field for the sake of novelty. Use Anchor when the sentence has to survive inspection.
Restrained, precise, confident. A still composition. One photograph, fully seen.

Illustrative evidence map
Highest-ranked cohort
FACT · sample
Illustrative only. In the real arena this row is a cohort returned by your run, with its call id.
Identity-checked seeds
FACT · sample
Illustrative only. References that fail resolution are shown as unresolved, not guessed.
Rejected world
UNKNOWN
Illustrative only. Insufficient worlds stay visible with their reason instead of disappearing.
What-if lab · design preview
Moving a slider computes an illustrative index with the published weights below. It does not score a film, does not call a provider and does not write an experiment. The real lab reorders existing candidates from a completed run.
Suggested anchor · selected anchor (matching). Evidence strictness outweighs exploratory range in this demonstration, so the strongest supported direction leads.
anchor = 0.5·E + 0.3·(100−R) + 0.2·(100−D) · bridge = 0.4·D + 0.4·E + 0.2·R · frontier = 0.5·R + 0.3·D + 0.2·(100−E)
Demonstration rule with published weights — not a production ranking model.
Open the evidence-backed arena ↗Enter a room

01 · Creator Studio
Brief, fingerprint, Cultural Multiverse, Test Card, archive. For one film and the worlds around it.

02 · Cinema Mode
Venue, city, programme. Curatorial notes, not a streaming wall. Geography only when the data exists.
03 · FilmSponsor
Live brand discovery when configured. No substitute brands. Approval after source review. Outreach is observed, not predicted.
Open FilmSponsor