I need a proof-of-concept “tender-scraper assistant” The flow is straightforward: • Scrape the main public tender portals (think CPPP, GeM, and similar) in real time, collect every PDF link on each results page, and download the files locally. • Run each PDF through an LLM pipeline—OpenAI, Claude, or any comparable model is fine—to extract the essentials: scope of work, estimated value, deadlines, required documentation, and anything else you can reliably surface. • Compare those extracted details against a single company-profile PDF I’ll supply and generate an auto-score that tells me, at a glance, whether it’s worth our bid effort. • Push the structured data to an .xlsx file and simultaneously create a concise markdown report for my team. Once the POC is validated, I also want the whole flow scheduled twice a day with automatic email or Slack alerts. I’m happy for you to lean on low-code AI tools where it speeds delivery, but lightweight custom code for the scraper and PDF processing is expected. Clean, well-commented source files, installation notes, and a brief walk-through video will be the key deliverables.