Sondo is
- Cultural affinity research grounded in Qloo results
- What over-indexes for your audience versus a baseline
- Hypotheses with a concrete plan to test them
Sondo turns a brand brief into evidence-backed campaign strategies, grounded in Qloo's cultural graph. Every cultural claim traces to a specific Qloo result.
Grounded in Qloo's cultural graph. Every claim cited.
Stored runs of the fictional evaluation briefs. Each opens instantly, with every claim linked to the Qloo result it cites.
You give Sondo a brand, an audience, a market and a goal. It researches Qloo, critiques what came back, and drafts three directions. Every cultural claim cites the result behind it.
Your brand, audience, market and goal. Sondo shows which parts Qloo can model, and which it can't.
Live Qloo queries across music, film, TV, podcasts, brands and places, for your audience and for a baseline.
Deterministic checks for coverage, empty categories and signals that are merely mainstream.
Three differentiated directions. Every cultural claim cites the Qloo result behind it, with a plan to validate it.
Cultural affinity, not purchase intent. The same boundary is printed on every report.
Short answers from the product notes.
Sondo is a cultural-intelligence strategist for brand and agency teams. You give it a brief: your brand, your audience, your market and a few things that audience already loves. It researches what over-indexes for that audience across music, film, TV, podcasts, books, brands and places, then writes three different campaign concepts. Every claim links to the evidence behind it, and creative bets are labelled as hypotheses.
From live Qloo requests during a run. Affinity is Qloo's relevance score for that query. It is not a probability, an audience share, or a count of people. Response bodies are not committed to the repo. Test fixtures are synthetic, with fictional names and fake ids.
No. It does not predict purchase intent, conversion, ROI, audience size or share, and it does not promise a campaign will work. A strategy is a hypothesis with a plan to test it. The validator blocks invented numbers and performance claims.
Audience lift runs each category twice: once with the audience signals, once as a location baseline. Comparing them shows what over-indexes for this audience. Lift is a Sondo heuristic, labelled as such. It is not a Qloo metric and it is not audience size.
The project is open source under the MIT License. This build has no accounts and no billing. Live Qloo calls need a hackathon API key and the agent needs a model key. Both stay on the server.
No login. Open a featured report to see a finished run, or build a brief and watch the agent research it live. A run takes about two to three minutes and keeps going if you leave; its report link works once it finishes. Each visitor gets a few live runs a day to protect the shared Qloo quota. A run marked partial could not ground all three concepts, and the report says why.
Bring a brief. Leave with citations.
New York City · Launch
CompleteStride Loop
Top findingKode With Klossy appears only for this audience, not in the market baseline (brand).
Austin · Launch
CompleteOat Hour
Top findingKEEN Footwear ranks higher for this audience than in the market baseline (brand).
Lisbon · Market entry
CompleteCasa Azulejo
Top findingWANT Les Essentiels appears only for this audience, not in the market baseline (brand).
A concept is a hypothesis with the Qloo results that suggested it, and a way to test it. The card below is copied from a stored report, not drawn for this page.
The brief stays in charge. Sondo researches your audience against a market baseline, then drafts three different angles.
From the featured Stride Loop report · New York City
Stride Loop steps into the everyday soundtrack of young New Yorkers, aligning with the chillwave, female‑voiced Lo‑Fi beats that define their urban vibe.
The top taste themes for this audience are chillwave, female singer, and Lo‑Fi Beat, signalling a shared sonic identity that can anchor a launch.
Built for the Qloo Agentic Hackathon 2026. The agent cites ledger ids. It does not invent scores, and it does not predict ROI.
Ledger rows behind that claim · Stride Loop report
Sondo runs each category twice. Once with the audience signals, once as a market baseline. The comparison shows what over-indexes here. Lift is a Sondo heuristic, not audience size and not a sales forecast.
The Hate U Give
Rank #3 vs #10
Rank in the audience query, then in the market baseline. Sondo heuristic.