l0ng-ai/papr

A fast, native RSS reader — plus an agent-facing CLI so your AI agent can read, search and triage your feeds from the shell.

Papr – A fast native RSS reader with an AI‑agent‑friendly CLI

What it is – Papr is a desktop RSS reader that stores everything locally and a companion command‑line tool (papr) that lets autonomous agents (Claude Code, Codex, OpenCode, etc.) read, search, and manage your feeds without a GUI.

Key components

  • Desktop app – native, offline‑first client with folders, tags, rules, full‑text extraction, audio playback, and built‑in AI helpers (summaries, Q&A, digests). It can sync read state with a FreshRSS server and is localized in English, Japanese and Simplified Chinese.
  • papr CLI – designed for agents. It exposes commands to list unread items, read full text, search, mark articles, refresh feeds, manage subscriptions, import/export OPML, and sync with FreshRSS/Miniflux. Output uses the compact TOON format (~40 % fewer tokens than JSON) and returns structured exit codes, making it easy for LLM‑driven agents to parse.
  • Shared core – both front‑ends use the same papr-core Rust crate, guaranteeing identical data handling.

AI integration – The desktop app can call any LLM API you provide to generate article summaries, answer questions about an article, or create periodic digests. The CLI can be invoked by an agent through the bundled papr-rss skill, which can be added with Vercel’s skills tool and even set up an ambient hook that notifies the agent of new feeds at session start.

Installation

  • Desktop – macOS via Homebrew Cask (brew install --cask l0ng-ai/papr/papr), Windows .msi, Linux AppImage or .deb. All binaries are on the latest GitHub release and macOS builds are notarized.
  • CLI – macOS/Linux via Homebrew (brew install l0ng-ai/papr/papr-cli), Windows zip download, or any platform by extracting the pre‑built tarball from the releases page. You can also build from source with Cargo (cargo build --release -p papr-cli).

Why it matters – Papr bridges the gap between personal information consumption and autonomous AI agents, giving agents a reliable, token‑efficient interface to stay up‑to‑date with your RSS feeds and act on them (e.g., summarizing new articles, triaging unread items, or feeding content into other workflows).