riszotto — Research from Your Terminal

Everything you need

A complete toolkit for working with your research library from the command line

🔍

Full-text Search

Search across titles, authors, years, and abstracts. Filter by tags, date ranges, and item types.

🧠

Semantic Search

Natural language queries powered by ChromaDB embeddings. Find papers by meaning, not just keywords.

📖

Read as Markdown

Convert PDFs to clean markdown. Paginate, search within, extract figures and tables.

📋

BibTeX Export

Export citations with configurable field filtering. Ready for LaTeX, Typst, or any reference manager.

👥

Group Libraries

Access shared Zotero group collections. Search across all libraries at once with --all-libraries.

🤖

AI Agent Ready

Built-in Claude Code skill for paper research workflows. Let agents search and read your library.

Install in seconds

Requires Zotero desktop running with the local API enabled

$ uv tool install riszotto
$ riszotto --help

# With semantic search
$ uv tool install "riszotto[semantic]"

$ pip install riszotto
$ riszotto --help

# With semantic search
$ pip install "riszotto[semantic]"

# Install as a Claude Code skill
$ npx skills add https://github.com/pythonfz/riszotto

# Then ask Claude to search your library

Your research assistant

Install as a Claude Code skill and let AI agents search your library

>_ claude code — ~/research-project

You

Find recent papers on graph neural networks for molecular property prediction in my Zotero library

Claude

I'll search your Zotero library for relevant papers.

Bash

$ riszotto search --semantic "graph neural networks molecular property prediction"

Key	Title	Author	Year
XK92MN	SchNet: A Deep Learning Architecture…	Schütt K.	2018
PL47FD	DimeNet++: Fast Directional Message…	Gasteiger J.	2020
QR83BV	MACE: Higher Order Equivariant…	Batatia I.	2022

Claude

Found 3 relevant papers. Let me read the most recent one.

Bash

$ riszotto show QR83BV --page 1

# MACE: Higher Order Equivariant Message Passing
## Abstract
We introduce MACE, a higher order equivariant message
passing neural network for fast and accurate force fields.
MACE achieves state-of-the-art accuracy on molecular
property benchmarks while maintaining computational…

Claude

Here's what I found:

MACE (Batatia 2022) introduces higher-order equivariant message passing for molecular force fields, achieving state-of-the-art on property prediction benchmarks.

DimeNet++ (Gasteiger 2020) improves directional message passing with faster training and better accuracy on QM9.

SchNet (Schütt 2018) is the foundational continuous-filter conv architecture for quantum chemistry.

Want me to export the BibTeX for all three, or read any of them in detail?

You

Export bibtex for all three