What I'd build for you
The demo above is the core. The production tool wraps it for your real archive:
How the production version works — and the exact stack
The demo runs a real search-and-extract on a synthetic CSV so it loads instantly. For your archive, the same pipeline reads your actual files:
Packaging
Delivered as a one-double-click Windows app (PyInstaller) or a local Streamlit page — your choice. No install headaches, nothing to host.
The synthetic archive this demo searches
⬇ Download the 2,547-row archive.csv the tool finds the newest row per part number, exactly as it would across your real files.
This is the tool from your brief — already working.
Send me your example quote and I’ll match your template exactly. Happy to do a quick screen-share first.