Tender PDF Bot & Checklist -- 2

I need a small automation that regularly visits government and corporate tender sites, searches for new PDFs using my own industry-specific and project-specific keywords, downloads each matching file, and then feeds the document straight into an LLM to extract the information I care about. The model’s sole job is to return a clear compliance checklist, nothing else.

Ideally the script will run headless, keep a simple log of URLs and timestamps, and save both the raw PDFs and the generated checklists (CSV or JSON is fine) in neatly named folders so I can audit later. Please write clean, well-commented code—Python with Selenium/Playwright plus an OpenAI or similar API would suit me, but I’m open to anything equally reliable.

Deliverables
• Source code and brief setup guide
• Configuration file or interface where I can edit keyword lists and add new target sites over time
• Sample run demonstrating the compliance checklist output from at least one government and one corporate tender PDF

Let me know the estimated turnaround and any clarifications you need.