HumanSourcer

HumanSourcer — Daily-refreshed pay-intelligence index for the AI-training and data-annotation gig economy: 786 distinct advertised hourly rates across 26 paying platforms (median $56/hr, recomputed daily) drawn from 7,402 live listings scraped and classified every 24 hours, spanning 74 tracked provider networks tiered into 16 expert networks, 51 gig/task platforms and 10 managed workforces. Each rate carries per-figure provenance (structured field vs rule-parsed vs model-extracted from listing prose, and whether the source hedged the number), a work-domain tag (coding, linguistics/translation…

HumanSourcer offers (Alternative, Sentiment, ESG, Fundamental) — 7,402 active listings indexed daily from 33 of 74 tracked networks; 786 distinct advertised rates across 26 platforms; per-provider daily count tables by use-case category with a visible continuous daily series back to at least 2026-08-17..

Direct cost input for AI-lab workforce budgeting and vendor selection (this is the only public window into what expert-tier evaluation work actually pays); benchmarking of in-house annotation rates against market; gig-economy wage and labour-exploitation research (the vendor's own finding that 1 in 15 advertised AI-training rates is under $10/hr); supply-chain diligence on AI-lab outsourcing (which network feeds which model programme, with ownership mapped); and, for AI buyers, a rare labelled corpus at the intersection of task taxonomy and compensation — supervision data for models that price or route human-data work, plus ground truth for RLHF/evaluation-workflow cost models that no lab publishes.

Coverage spans US; Human Capital & Employment Services, Application Software; alternative, sentiment, ESG, fundamental; equities.