Ben Softie (evm/spice 🌶,🐝)

Your agent decides. Something smarter should execute. Cooking this on @strikeperps ecosystem

Introducing Crowdline. Before your shielded ZEC leaves Zcash, it shows you what the swap reveals, scores it, and fixes it: confidential quote, a crowded amount, a one-time refund key, paid from your shielded balance. No wallet connect. No custody. crowdl.site

Where should an indie artist actually tour? Ask an AI and you get the biggest cities in the country. Small teams make the same guess, and that’s how tours lose money: big rooms where nobody knows the act, while the towns where the fans really are get skipped. So for the @Qloo Agentic Hackathon I built Opener: an AI tour-routing agent that plans the tour around where Qloo’s taste graph says the audience is. The test that convinced me I asked Gemini, with no data, to plan a North America tour for Phoebe Bridgers. It picked Los Angeles, New York, Chicago and San Francisco. None of the four match @Qloo top heat cells for her audience. Opener’s route starts in Boston (#1 in Qloo’s US heatmap), Vancouver and Montreal. A plain “biggest cities” list did no better: five of its eight picks had no @Qloo match at all. For stars whose fans are everywhere, like Sia, the guess and Qloo mostly agree (6 of 8). That’s the point: Opener matters most for indie acts, the ones who can least afford a wrong city. What it does You type an artist and a region. Opener: • resolves the artist in Qloo (and asks you to pick if the name is ambiguous) • maps where the audience over-indexes in each country • turns that into a routed 5–8 stop tour • explains every stop in plain words: its rank in Qloo’s heatmap, its affinity score, how many heat cells it covers • suggests openers per city from artists Qloo recommends to the same audience in that city • shows who the crowd is and what else they like: brands, shows, podcasts, local places Every plan has a “Generic vs Opener” button that puts the no-data guess next to the Qloo -ranked route, so you can see the difference for yourself. How it works A Gemini agent decides which Qloo calls to make: heatmaps per country, recommendations, audience comparisons, demographics, trends. Code does everything numeric: it maps heat cells to cities, ranks them within each country and routes one trip. The model only writes the words, and the server throws out any city or opener Qloo didn’t actually return. An “Under the hood” drawer lists every @Qloo call behind the plan, so every number can be checked. What I learned Read the raw numbers before trusting a ranking. My first worldwide plan scored the top city in every country as equally strong. @Qloo heatmap scores turned out to be relative within each country, so now Opener ranks cities within each country. And treat a model’s answer as a draft to check, not an answer to show. Demo: youtu.be/-qubntGJhiE Code: github.com/softalpha0/Opener