ArXistant documentation

Your personalized astronomy reading assistant

Install ArXistant, build a useful saved-paper library, and understand how its local ranking model turns new arXiv submissions into a focused daily reading list.

Start here

01

Installation

Set up the Chrome extension and local server on macOS, Debian/Ubuntu, Windows, or another Linux system.

02

User guide

Work through daily papers, reminders, saved items and tags, the chat reading helper, search, publication import, and model controls.

03

How it works

Explore the local architecture, ranking pipeline, training lifecycle, data storage, and platform startup.

What stays local

ArXistant runs a small server on your computer. Your saved-paper database, ranking model, manual keywords, notes, highlights, and generated pages remain in its local data directory. Network requests go directly to arXiv and, when configured, NASA ADS; if you use the Chat page, questions also go to the LLM provider you choose (a local Ollama keeps even that on your machine).

Current platform support

Platform Installation Automatic server startup
macOS Repository checkout Bundled helper application
Debian/Ubuntu Experimental .deb package systemd user service
Other Linux Manual Python setup User-configured
Windows Manual Python setup Not yet available

Ready to begin? Continue to the installation guide.