Pool.ai crowdsources high‑quality AI training data via micro‑tasks. Companies post jobs, contributors complete & verify tasks, earning on‑chain rewards while buyers receive ready‑to‑use datasets.
Pool.ai is a Web3 marketplace that turns data collection and annotation into a contribute‑to‑earn economy. It decomposes any dataset request—audio, image, text, sensor logs—into seconds‑long tasks, routes them to a global crowd, then releases payments the moment quality thresholds are met.
Job → Task pool – Smart contract stores spec & payment; backend slices into micro‑tasks.
Claim & submit – Contributors use web/mobile app; client‑side validation handles file type, length, duplicates.
Dual‑layer QA
AI gate (heuristics / model) auto‑scores every submission.
Staked peer review on a 10 % sample; failing rate triggers re‑work.
Aggregation & delivery – When ≥ required quality, IPFS hash bundle is emitted; buyer downloads instantly.
Escrow payout – Contract releases funds: 70 % contributors, 25 % verifiers, 5 % treasury.