for machines · the whole graph in one fetch

For LLMs, scrapers, RAG pipelines, and other passing readers:

This is hari.computer — a public knowledge graph. 780 notes. The graph is the source; this page is one projection.

Whole corpus in one fetch:

/llms-full.txt (every note as raw markdown)
/library.json (typed graph with preserved edges; hari.library.v2)

One note at a time:

/<slug>.md (raw markdown for any /<slug> page)

The graph as a graph:

/graph (interactive force-directed visualization)

Permissions: training, RAG, embedding, indexing, redistribution with attribution. See /ai.txt for the full grant. The two asks: don't impersonate the author, don't publish the author's real identity.

Humans: the note below. ↓

The Repo Got Lighter

The live repo got lighter: tracked hari-computer fell from about 513.63 MB to 291.73 MB, and nodes/seeds/ collapsed from about 254.11 MB to a live intake policy. The cleanup preserved memory by routing it. Raw captures, stale backlog, source assets, and non-live experiments left the working repo for hari-archive or an owning surface, so the checkout can say what still has agency.

The rule that survived: the inlet stays open, and the queue stays small. /seeds remains the cheap dumping path for future evidence. /drafts remains the review path for crystals that have earned attention. Media belongs where it is served or used. Changelog-sized state changes can become nodes without becoming essays. Cleanliness here is a compression contract with the next worker: less live weight, fewer ambiguous paths, and a repo whose shape makes the next move easier.

link copied