AI programming director for independent venues · powered by Qloo's taste graph
Your crowd already told the world what they love.
Marquee turns it into a sold-out month.
Ask a generic AI how to fill a quiet Tuesday and you get "trivia night", the same answer for every venue on earth. Marquee reads the cultural taste of the people who already love your room, tests genres against them, and programmes nights they over-index on, each with its Qloo evidence and a draw forecast.
- 9 Qloo-backed tools
- 6 taste domains
- Every night explained
- Draw & ROI forecast
- Compare any two venues
- MCP server
Agent at work
Who your crowd is
Age skew vs. the general population (Qloo demographics). Right = over-represented.
Genre radar
Each bar = mean Qloo affinity of the top 5 entities in that genre for this crowd. Marquee tested these live; the month uses the strongest.
Your month, programmed
Open Why this works on any night to see the Qloo signal, its rank and the evidence behind it.
Draw & ROI forecast
planning estimateEach night's draw index (0–100) comes only from Qloo signals: 50% crowd fit (affinity), 25% genre strength, 25% reach (popularity), +5 for audience growth, +5 for a bridge pick. Enter your room's numbers to turn it into guests and money. Expected guests = quiet-night guests + (capacity − quiet-night guests) × draw × conversion. Net = extra guests × (ticket + bar spend) − cost of the night.
Gut-check an idea
Type any genre or format. Marquee finds it in Qloo and scores it against every genre it tested for this crowd.
Grow without losing your regulars
Where to put the flyers
Neighbourhoods where this crowd's taste concentrates (Qloo heatmap). Dots = taste hotspots; ★ = your venue.
Partners next door
Places within 1.5 km whose audiences overlap with yours.
Rooms competing for the same crowd
Ready-to-send pitches
Taste twins in other cities
Same kind of room, crowds that love the same things. Borrow their playbook, swap touring acts.
Brands your crowd over-indexes on
Provenance: every Qloo call this run made
Endpoint, parameters (never the key), latency and whether it came from Marquee's private cache.
| # | Endpoint | Parameters | ms | Cache |
|---|
Proof: a generic AI vs Marquee
The same LLM suggests nights with no taste data; Marquee then gut-checks each suggestion against this crowd in Qloo.
Compare two venues, in any two cities
multi-venueRun a second room, plan a tour, or scout a new city: Marquee scores the same 12 music genres against both crowds from Qloo and tells you where they agree and where they differ.
Use Marquee from any agent (MCP)
Marquee is also a public, stateless MCP server. Add it to Claude, Cursor or any MCP client and ask: "Plan next month for Smalls Jazz Club and gut-check a salsa night."
Tools: marquee_plan_month, marquee_gut_check, marquee_crowd_profile, marquee_compare_venues. Local stdio: node bin/marquee-mcp.js.