Weekly listings-to-table extract
Every Monday, turn messy listing pages — job boards, directories, product or event listings — into one clean, deduplicated table, extracted by ScrapeGraphAI.
Every Monday, extract fresh listings from these pages with ScrapeGraphAI and send me a table. Pages to extract from: - <https://example.com/jobs — listing pages like job boards, directories, marketplaces> - <add more listing URLs> The fields I want per listing: <e.g. "title, company, location, salary if shown, posted date, link">. What counts as relevant: <describe your filter — e.g. "remote roles only, senior level">. 1. Use ScrapeGraphAI's extract tool on each page to pull every listing as structured records with exactly those fields. These pages often render client-side, so don't fall back to a plain fetch. 2. Filter to relevant entries and drop anything you already reported in previous weeks — I only want listings that are new since the last run. 3. Present one table sorted newest-first, then a one-paragraph read on the week: how many were new, and anything notable. If nothing new appeared, say so in one line. Cap the run at 5 pages so credit spend stays predictable.
New listings — Mon 29 Jun 2026 7 new since last week (31 extracted, 24 already seen or filtered) Title Company Location Salary Posted ──────────────────────────────────────────────────────────────────────────── Senior Platform Engineer Driftwave Remote $175–205k 27 Jun Staff Data Engineer Nimbus Labs Remote EU €120–140k 26 Jun Infra Lead Corely Remote US — 25 Jun …4 more The week: hiring picked up — 7 new senior infra roles vs 2 last week. Two explicitly mention on-call pay, which was rare in this board before.
Anyone who's tried to watch a job board, a supplier directory, or a marketplace knows the loop: open the page, scroll past everything you saw last week, try to remember what's actually new. This routine runs that loop for you — ScrapeGraphAI extracts every listing as structured rows, the agent drops what it already showed you, and Monday's email is only the new stuff, in a table.
How it works
- AI extraction, not scraping selectors: you describe the fields in plain English and
- ScrapeGraphAI pulls them from each listing — no CSS selectors to write or fix when the page
- changes.
- The routine remembers previous weeks, so the table is new listings only, already filtered
- to your criteria.
- Capped at 5 pages per run, so a scheduled Monday sweep spends a predictable amount of your
- ScrapeGraphAI credits.
Make it yours
- Swap in your real sources: niche job boards, RFP directories, grant listings, real-estate
- pages, conference CFPs.
- Change the fields to whatever your table needs — deadlines, prices, contact links.
- Ask for the table as a CSV in My Files too, and it accumulates into a dataset you can open in
- a spreadsheet.
Run “Weekly listings-to-table extract” on autopilot
Copy it to your account, tweak the details, and it runs mondays at 9:00 am.
Use this routine →More Research templates
Every Monday, a digest of newly published and trending Hugging Face models, datasets, and Spaces in the topics you follow, emailed and ready to skim.
A short, opinionated brief of what actually happened in your industry, in your inbox before work.