AI Bid Filter Automation

I need a small, self-contained bot that automatically reviews every bid I receive and pushes anything that looks like spam into a separate queue so I never even see it. The focus is purely on filtering out inappropriate bids; I’m not asking for scoring, ranking, or highlighting—just solid, reliable rejection of junk.

Here’s what the bot must do:

• Run keyword and pattern checks that reliably flag spam-like content, specifically bids that are (a) blatant copy-paste repeats, (b) stuffed with excessive links, or (c) obviously lifted word-for-word from other proposals.
• Work in real time (or on a short poll interval) so flagged posts disappear from my view before I open them.
• Expose a simple dashboard or JSON log that shows the bid, the rule it tripped, and the confidence score, so I can quickly skim false positives.
• Allow me to toggle individual rules on or off and retrain thresholds without touching the code—an editable config file or lightweight admin panel is fine.

Tech is up to you, but I’m already running Python, so sticking with something like spaCy, scikit-learn, or even an OpenAI moderation endpoint will slot neatly into my stack. Whatever you choose, ship clear instructions and a five-minute setup script so I can plug the bot into my existing workflow and start filtering immediately.

Acceptance criteria
1. A live demo on sample bids showing each of the three spam patterns being caught.
2. Zero critical false positives on a clean test set I’ll provide.
3. Delivery of commented source, requirements.txt, and a one-page README.

If you have quick turnaround expertise in NLP or rule-based text classification, this should be a straightforward build. I’m ready to get started as soon as you are.