New freelance jobs scraped, scored by fit, and surfaced daily so nothing good gets missed.
Scanning a job board by hand is slow and inconsistent. Good-fit work slips by while low-fit noise eats attention.
A scheduled scrape pulls fresh postings, an LLM rates each against a real ICP rubric (fit reasons, red flags, angle bullets), rows land in a CRM-shaped board, and only high-scoring jobs trigger a batched notification. Every posting is deduped against every prior row so nothing repeats.
The score-and-surface engine works for any inbound stream: RFPs, grants, marketplace listings, inbound leads.