About

How the system works, and why it is built this way.

Disaster Clippy grew out of a specific frustration: when you search for preparedness information you get a wide, inconsistent spread of results. Some are excellent. Some are outdated. Some are wrong. The system is designed to search a curated, known collection and show you exactly where each answer came from.

How sourced answers work

The model only works inside the current collection. It searches documents in that collection, synthesizes an answer, and points back to the source it used. AI can state things with confidence that are wrong - the fix here is to constrain what it knows and force it to show its work. Every answer in Disaster Clippy points back to the source it was drawn from. If the answer is wrong, you can read the original source yourself.

What the collection is

Narrow and legible beats wide and opaque. Open search returns whatever ranks. Disaster Clippy searches a fixed collection you can inspect. You can see what is being searched, choose which sources to include, and trust that nothing outside them is influencing the answer.

Why hosted and local both matter

A preparedness tool that requires internet access is not a preparedness tool. The hosted app is the easy starting point, but the local runtime runs the same search experience on your own machine, on a Raspberry Pi, or fully offline with no external APIs and no data leaving your hardware.

How the builder path works

Full control belongs with the people who understand their sources. Scraping, transcript generation, and pack building are deliberately kept in the local admin path. That keeps the public app clean while giving maintainers and contributors the full system for bringing their own data.