Open tool · astro-agalmic · v0.1
Take the knowledge architecture.
The discovery system built for Agalmic Research is now a reusable Astro integration: semantic identity, relationship validation, full-text search, visible knowledge history and generated context without a backend.
Agalmic by construction
Useful machinery should become cheaper for the next person to possess.
astro-agalmic is licensed under 0BSD. Anyone may use, copy, modify and distribute it for any purpose, with or without attribution or fee.
What it gives away
The pattern, not our vocabulary.
A site can preserve a large, changing knowledge space while keeping scarce attention focused on meaningful directions.
- Semantic registry: stable IDs, object kinds, topics, commitment states, relationships and canonical URLs.
- Build validation: broken relationships, duplicate canonical identities and rendered-title drift become build failures.
- Full-text search: a static search index is generated from the rendered site after Astro builds.
- Knowledge context: reusable Astro UI for origins, neighbours, handoffs and meaningful next directions.
- History: schemas and conventions for renamed, narrowed, merged and superseded objects.
- Content collections: optional schema factories support migration toward Astro's typed content layer.
The package deliberately does not require concepts such as “Memory Palace”, “Active Frontier” or “Agalmic Research”. Host sites choose their own language and visual design.
Minimal use
A few lines in Astro config.
import { defineConfig } from 'astro/config';
import agalmic from 'astro-agalmic';
export default defineConfig({
integrations: [
agalmic({
registry: 'src/data/knowledgeIndex.json',
history: 'src/data/knowledgeHistory.json',
handoffs: 'src/data/handoffs.json',
}),
],
});At build time the integration validates the semantic graph and generates search-index.json. Reusable Search and KnowledgeContext components can then project the same knowledge model into the reader interface.
Standards boundary
Glue, not another private ontology.
The integration does not replace W3C PROV, CRediT, RO-Crate, Astro content collections or specialist static search tools. It supplies the smaller missing layer between a pile of pages and a navigable research system.
If a mature standard already solves a deeper problem, reuse it. If a corpus eventually needs Pagefind or another dedicated search engine, replace the lexical index without changing the semantic architecture.
Release status
Source first.
Version 0.1 is an extracted release candidate and is already used by this site.
The source package is public now. The package manifest is prepared for npm publication, but the npm name and release should be treated as unclaimed until publication actually succeeds.