Map a real-world dataset to particle positions
Take any public dataset (CSV). Propose how to map columns to particle x/y/z/color/size. What patterns would become visible in 3D? Submit: dataset link, mapping spec, reasoning. 300+ words.
Choose a task, claim the bounty, and earn SPARE™ rewards. Every contribution is cryptographically verified.
Take any public dataset (CSV). Propose how to map columns to particle x/y/z/color/size. What patterns would become visible in 3D? Submit: dataset link, mapping spec, reasoning. 300+ words.
Translate the hero section + feature descriptions into any non-English language. Must be a native speaker. Submit original + translated text.
The Discovery Feed shows events like 'New arXiv paper found' and 'User correction absorbed.' Design the card that displays one event. Submit as HTML/CSS mockup or Figma link.
When Council is deliberating (4 models working), what should the user see? Design a 3-5 second loading state. Submit as CSS animation, Lottie, or detailed description + sketch.
Pick Chat, Council, Build, Evidence, Explore, or Earn. Write a user guide: what it does, how to use it, tips and tricks. Include screenshots.
Use Council mode. Find 5 questions where the 4 models give genuinely different answers. Document each: the question, what each model said, which answer seems best and why.
Create a mathematical equation that produces an interesting particle visualization. Submit: equation, parameter ranges, screenshot of the result, and explanation of what it shows.
Use the workspace for a real task. Write 200+ words about your experience. What worked? What surprised you? What would you change?
What do you currently use (ChatGPT, Claude, Gemini, Cursor, Copilot, etc)? Try the Hive for the same tasks. What's better? What's worse? What's different? 300+ words.
Define a complete color palette for a new VSE visualization mode. Submit: accent color, background color, particle colors (primary/secondary/tertiary), blend mode, and a name for the preset.
Pick 5 questions. Ask both ChatGPT and the Hive. Document: which gave better answers? Where did the Hive's multi-model approach help? Where didn't it? 300+ words.
Find 5 medical questions where PubMed fallback fired. Click each citation link. Verify the paper actually supports the claim. Submit verification report.
Describe a physics simulation you'd want to see in the particle engine. What does it show? What parameters should be adjustable? What would it teach? 200+ words.
Open /workspace on a phone. Screenshot each mode (Chat, Council, Build, Evidence, Explore, Earn). Report any layout issues, cut-off text, or broken buttons.
Open /workspace, run 10 questions in Council mode. Rate each consensus response: Correct / Partially Correct / Wrong. Submit ratings via correction button.
Use the workspace for at least 15 minutes. Then submit one feature request: what you wanted to do, why, and how you'd use it if it existed.
Visit /pricing in incognito. Verify all buttons say 'Free during beta.' Try to checkout. Confirm it's blocked. Report results.
Pick a project or consult the collective before choosing a bounty.