Preparedness collection
Mixed practical sources
Current examples include Appropedia, WikiHow, BuildItSolar, Kiwix ZIM archives, PDFs, and other resilience reference material packaged into one searchable library.
Collections
What is in the collection right now.
The hosted app currently uses a preparedness-focused collection assembled from practical reference sources, offline archives, and document libraries. Under the hood those are packaged as packs, but the public-facing idea is simpler: this is the library you are searching today, and the same model can be rebuilt for your own domain.
Visible scale
Named sources, not a black box
This is not a black-box dataset. The example deployment is built from real sources with clear collection boundaries. You can see what is in the library, what is being searched, and how that pattern could work with your own material.
Why it matters
Same engine, different data
HTML sites, PDFs, video transcripts, and ZIM archives all normalize into the same searchable pattern. The collection changes; the runtime stays stable.
The runtime and the library stay separate
Most search tools bundle their knowledge with the application. Disaster Clippy treats the collection as a first-class artifact that can be versioned, released, and distributed independently. Under the hood, packs are the portable unit. Publicly, what matters is that you can update the collection without rebuilding the app.
What is in the example deployment
The current demo collection is preparedness-oriented on purpose. It mixes household emergency guidance, practical DIY material, survival and resilience references, and offline-friendly archives. That makes the app feel real as a product while also showing developers the range of source types the system can absorb.
Kiwix ZIM archives are an especially useful bootstrap path. If you already trust a Kiwix library, you can use those archives as source material for a local Disaster Clippy deployment and start with content you already know.
Building your own collection
The public repo includes ingestion tooling for websites, PDFs, and video archives. Point it at a source, run the pipeline, and you get a collection the runtime can load. The same tools used to build the preparedness example are available for your own data.
The workflow is straightforward: choose a source, ingest it, package it, and run the same search interface against it. If you can picture your own library in that flow, that is the point. Learn how to do that here.