Use Cases
01Quick Answer
Use scrape_with_actions (5 credits) to click "load more", scroll, or sign in so the listings actually render, then extract_structured (3 credits) to map each posting onto your own JSON schema -- title, location, salary, remote flag. About 8 credits per board page, and one schema covers boards with different markup.
02The brief
Job boards hide their listings behind "load more" buttons, infinite scroll, and login walls, so a plain HTTP fetch returns an empty shell. Every board also names its fields differently, which breaks selector-based scrapers on the next redesign.
CrawlForge scrape_with_actions drives a real browser -- click, type, scroll, wait -- so the listings render before extraction, and extract_structured maps whatever markup a board uses onto one JSON schema you define.
03In code
1// Load a paginated board, then extract each posting as JSON2const board = await mcp.scrape_with_actions({3 url: "https://boards.example.com/jobs?team=engineering",4 actions: [5 { type: "wait", selector: ".job-card", condition: "visible" },6 { type: "click", selector: "button.load-more", clickCount: 3 },7 { type: "scroll", direction: "down", distance: 4000 },8 ],9 formats: ["json", "markdown"],10});11 12// One schema, any board layout13const posting = await mcp.extract_structured({14 url: "https://boards.example.com/jobs/senior-backend-engineer",15 schema: {16 type: "object",17 properties: {18 title: { type: "string" },19 location: { type: "string" },20 salaryRange: { type: "string" },21 remote: { type: "boolean" },22 postedAt: { type: "string" },23 },24 required: ["title"],25 },26});27 28console.log(posting.data);04The pipeline
05Questions
04
Use scrape_with_actions to drive a real browser: wait for the listing selector, click the load-more button, and scroll before extraction runs. The postings render first, so you get real content instead of an empty shell.
Yes. scrape_with_actions supports type and click actions plus formAutoFill, so it can submit credentials and scrape the page that follows. Only sign in to accounts you own or are authorized to access.
Define one JSON schema — title, location, salary range, remote flag, posted date — and pass it to extract_structured. It maps each board's markup onto your shape, and falls back to CSS selectors when no LLM is configured.
About 8 credits covers a board page: 5 for the scrape_with_actions browser pass and 3 for each extract_structured call on a posting. Use batch_scrape when you are pulling hundreds of detail pages a day.
06Keep exploring
Enrich sales leads with company data, tech stacks, and contact information from the web. Turn a bare domain list into a qualified, sales-ready pipeline.
Track competitor prices and product availability across hundreds of e-commerce sites. Detect undercutting the day it happens and reprice with confidence.
Start forging
Every new account gets 1,000 free credits. No credit card required.