A Python library that gives AI agents a reliable, deterministic interface to build and maintain a research library. Two layers: a reference store for epistemic objects, and a knowledge layer of synthesis notes, concept pages, and reading paths derived from them. Local-first, no GUI, no human workflows.
Runs in your terminal, alongside your coding agent. One install command, one init command.
Install with uv — not on PyPI yet, so this pulls straight from GitHub
uv tool install git+https://github.com/abrahambahez/loretoolsor download a prebuilt binary (no Python required) →
Point it at a library — required once per machine, nothing is created automatically
lore init /path/to/library
Optional: give your agent the CLI reference and the reading
workflows — download loretools-skills-en-vX.Y.Z.zip from the
Releases page
and unzip it into ~/.claude/skills/. Includes loretools-references for
the CLI, plus structural-reading and analytical-reading for mapping and
close-reading a text once it's in the library.
Research workflow — core
lore extract paper.pdf
lore lookup '{"DOI":"10.1000/xyz"}'
lore staging stage '<json>' --file paper.pdf
lore staging merge
lore refs filter --query "boundary objects" --year 2019
lore files attach vaswani2017 paper.pdf
Envisioned — not built yet
We envision a transition from isolated reference metadata to shared epistemic objects that enable deeper conversations.