PaperPulse

PaperPulse is an Apple-platform app for research reading on iPhone and Mac. It aims to turn the daily paper feed into a complete local workflow: scheduled discovery, authority-aware filtering, open-access PDF downloading, generated summaries, offline reading, and library management.

The public repository currently documents the project direction and implementation plan; the local source code is not pushed there yet.

Insight. The most useful idea in the README is that a paper assistant should be careful before it is clever. PaperPulse puts deterministic academic sources first, downloads only open-access PDFs, keeps rule-based screening ahead of LLM reranking, and asks the user to bring their own model provider configuration. That makes the app closer to a trustworthy research desk than a generic summarizer.

Highlights.

• Lets users configure research fields, keywords, exclusion terms, daily paper limits, and institution, journal, or conference preferences.
• Plans to rely on academic APIs such as arXiv, OpenAlex, Crossref, and Unpaywall before any model-generated reasoning.
• Treats paywalls and subscriptions as hard boundaries by downloading only openly available PDFs.
• Uses user-provided LLM API keys with OpenAI-compatible providers, custom base URLs, and model capability settings.
• Targets independent iOS and macOS experiences for refresh, downloading, offline reading, notifications, settings, and local library management.
• Tracks a SwiftPM-based core with models, pipeline protocols, source adapters, ranking logic, PDF handling, provider abstraction, and tests described in the README.

Repository. GabrielMu2006/PaperPulse