Keeper of the Chronicle
"The spellbook for your codebase — chronicle decisions, context,
and lessons your AI companions can actually read."
AI coding tools are stateless. They don't remember why you chose PostgreSQL over SQLite, that the auth layer must never bypass JWT validation, or that the frontend team deprecated the v1 API six months ago. You end up re-explaining the same context in every session.
Lore fixes that. It is a local AI memory system for software projects. You capture knowledge once as plain YAML alongside your code — then Lore publishes it as instruction files that GitHub Copilot, Claude, Cursor, Codex, and other AI tools read automatically.
By default, Lore exports CHRONICLE.md plus lean adapter files so the same project description, security preamble, and memory entry points stay consistent everywhere.
No external database. No API keys. No cloud sync. Everything lives in .lore/ next to your code.
A single piece of knowledge: a decision, a fact, a hard-won lesson. Short, specific, retrievable by semantic search.
A named collection of spells. Default tomes: decisions, facts, preferences, summaries. Add your own in .lore/config.yaml.
A raw artifact saved as-is for later distillation — a pasted session log, a git diff, a Slack thread. It lands in .lore/relics/ untouched until you're ready to extract spells from it.
New users should start with onboarding. Lore walks through the store layout, security policy, your first spell, and publishing in one guided ritual.
lore setup semantic if you want embeddings beyond TF-IDF fallback.Ready to begin? Here are the spells to summon Lore into your realm.
Lore ships a retro phosphor-green terminal browser — a full interactive TUI for browsing, searching, writing, and exporting your spellbook without leaving the command line.
x on a selected spell to preview/apply related links.lore/ on disk and updates instantlyOpen it in a split pane while you code — your full project memory, always one glance away.
When things are moving too fast to curate, capture the raw artifact first. Relics preserve noisy context as-is, then let you distill the verified parts into proper spells later.
Paste notes, import a file, snapshot a git diff, read from clipboard, or pipe in text from another tool.
Walk the artifact spell by spell, choose the right tome, and keep a traceable link back to the source relic.
Lore is designed for layered trust. The chronicle should be reviewed, reliable memory. Raw intake belongs in relics first.
CHRONICLE.md when a fact or decision is reviewed and ready for everyone to inherit.lore relic capture, then distill only verified points into spells.verified, needs-review, or deprecated to express trust clearly.lore trust refresh and explain one score with lore trust explain <id>.Install a post-commit hook that can extract knowledge from commits and regenerate all export files automatically.
Watch .lore/ for changes and export the moment the memory store changes.
Verify and repair store health with machine-readable JSON output. Run health checks in strict mode, set model timeouts, or apply automatic fixes.
Run a clean isolated CLI smoke test before publishing to catch packaging and command regressions.
Harmonize scans your spellbook in one pass and does two things: groups closely related spells into candidate rollup summaries, and flags entries that likely contradict each other. By default it is report-only — nothing is written until you say so.
Run a dry-run to see rollup candidates and contradictions without touching the store.
Persist a single rolling snapshot summary. Each apply updates it in place — no accumulated clutter.
Run harmonize automatically in the background daemon whenever the store changes.
Optionally ask an OpenAI-compatible model to write the snapshot text. Falls back silently to the local deterministic summary if no key is set.
Lore commands now emit machine-readable JSON for scripting and CI pipelines. Check health, lint, search, and list operate in strict mode. Auto-repair broken stores without manual intervention.
Gate deployments on store health. Use compact JSON and strict mode for fail-fast CI workflows.
Get structured JSON results for programmatic search and filtering.
Lint and list also support JSON. Combine with jq or other tools for custom assertions.
Detect problems and apply fixes automatically. Preview changes with --fix-dry-run before committing.
Lore uses a two-layer export model: one canonical chronicle, then thin instruction files that point back to it. That keeps token overhead low while preserving full project memory when needed.
/lore prompt in Copilot Chat to pull chronicle context on demand.
.lore/config.yaml.
lore onboard, lore init, lore add, lore edit, lore associate, lore list, lore list --json, lore remove
lore search <query>, lore search <query> --json, lore index rebuild, lore setup semantic, lore setup extract-patterns
lore relic capture, lore relic list, lore relic view, lore relic distill
lore export, lore export --format prompt, lore config <key> <value>
lore security, lore lint, lore lint --json, lore trust refresh, lore trust explain <id>
lore harmonize, lore harmonize --apply, lore harmonize --apply-resolutions, lore setup harmonize
lore hook install, lore awaken, lore awaken --harmonize, lore slumber, lore sync, lore doctor, lore ui
lore associate-audit, lore associate-heal, lore associate-prune, lore associate-relink, lore associate-repair