Paperbrace is a local-first literature navigator for a researcher’s PDF library (e.g., a Zotero attachments folder).
Disclaimer: Paperbrace is experimental, work in progress software. Use at your own risk. It comes with no guarantees of any kind, including correctness, security or suitability for production. See the LICENSE and docs/privacy.md for more information.
- Index a local folder of PDFs and enable QA/navigation over the user’s library
- Ground all answers in retrieved passages with page-level citations (“cite-or-refuse”)
- Run locally by default (no uploads, no external services)
Ask things like:
- “Which paper mentions X?”
- “Which paper claims improvement over Y?”
- “Summarize paper Z’s contributions and limitations.”
- “Compare A vs B on assumptions/datasets/metrics.”
Local-first by design: everything runs on your machine. No uploads, no servers.
- Running user-provided code
- Opening/executing links
- “Write my paper” mode (this is for reading/understanding/navigation)
- Hosted service operated by us
L0 (Search + Cite MVP).
- Index a folder of PDFs into a local SQLite DB
- List sources
- Extract per-page text into the DB
- macOS / Linux
- Python (project-local via
pyenv localrecommended) - SQLite (bundled with Python)
From repo root:
python3 -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install -e .
```bash
# Index PDFs:
paperbrace index --pdf-dir /path/to/pdfs --db paperbrace.db
# List:
paperbrace list --db paperbrace.db
# Extract:
paperbrace extract --paper-id 1 --db paperbrace.dbMIT (see LICENSE)
Attribution is not required. If you use Paperbrace in research or a public project and decide to give credit, you can credit it as: Keyvan Rahmani, Paperbrace (2026).