I cannot access the LICENSE file directly via the provided URL. However, the badge in the README indicates a license is present. I will proceed assuming a standard open-source license, but this is a point for potential clarification.

Paper Pilot

Paper Pilot is a local-first desktop research assistant designed for scientists to manage academic papers efficiently. It addresses the common pain point of juggling multiple browser tabs for paper discovery, reading, and synthesis. By running entirely on the user's machine, it offers a private, secure, and cost-free alternative to cloud-based solutions, ensuring no data leaves the user's computer.

How It Works

The system functions as an AI co-pilot for research workflows. Upon receiving a research query, Paper Pilot simultaneously queries over eight academic databases. Downloaded papers are converted into an AI-ready text format (MarkItDown) and then indexed using SQLite with FTS5 for full-text search and sqlite-vec for vector similarity search. An AI agent, capable of running locally via Ollama or through OpenAI-compatible APIs, synthesizes insights from this local corpus, grounding its responses in the retrieved papers with citations.

Quick Start & Requirements

• Primary install: Clone the repository, navigate to the directory, and run npm install.
• Requirements: Node.js >= 22.18.0, Python 3.11+. Ollama is optional for local AI inference.
• Running: Use npm run dev for development or npm run build followed by npm run package for a desktop application. Note: The development server defaults to http://127.0.0.1:5173.

Highlighted Details

• Multi-Source Paper Crawling: Integrates with 8+ academic sources including OpenAlex, Crossref, Semantic Scholar, PubMed/PMC, arXiv, Europe PMC, CORE, and Unpaywall, with experimental Google Scholar support.
• AI-Assisted Synthesis: Features local Ollama support for offline AI, OpenAI-compatible API integration, and a tool-calling agent with citation grounding within a chat workspace.
• Smart Local Storage: Utilizes SQLite with FTS5 and sqlite-vec for efficient full-text and vector search, with project-scoped storage to maintain research context separation.
• PDF → Knowledge Pipeline: Includes automatic open-access PDF fetching, MarkItDown conversion, and Python virtual environment isolation for scripting tools.

Maintenance & Community

The project is described as "Built by a researcher, for researchers." Specific details on community channels (like Discord/Slack) or a formal roadmap are not explicitly provided in the README.

Licensing & Compatibility

A license is indicated via a GitHub badge, suggesting an open-source license. Further clarification on the specific license type and any associated restrictions for commercial use or closed-source linking would be beneficial.

Limitations & Caveats

This project is an early v1, currently undergoing hardening for production reliability. The Google Scholar crawler is experimental, and packaging for macOS/Linux has not yet been tested. Cloud synchronization and collaboration features are not planned.