Skip to content
pkm

Search & Heat

pkm builds a SQLite index from Markdown files. The index powers full-text search, backlinks, and heat scoring.

Section titled “Search”

Search uses SQLite FTS5 and includes note titles, bodies, tags, and aliases.

Terminal window
pkm search kafka
pkm search --tags architecture
pkm search "data pipeline" --type dag --scope project

Filters use AND logic for tags.

Heat

Section titled “Heat”

Heat is a relatedness score between notes. It combines content similarity, shared tags, direct links, and co-references.

Terminal window
pkm heat compute
pkm heat show "Architecture Decision: Switch to Kafka"

Signals

Section titled “Signals”
SignalWhy it matters
Content similarityNotes discuss similar terms
Shared tagsNotes are classified together
Direct backlinksNotes explicitly reference each other
Co-referencesNotes point to the same third note

The score is not a truth metric. It is a discovery tool for surfacing nearby context.