AI 摘要与问答需自备 API key；CLI 在 Windows 需手动解压并把 papr.exe 加入 PATH；Agent 主动感知订阅需额外运行 papr setup 配置 SessionStart hook。
papr
l0ng-aiMIT活跃维护用 Rust 写的原生 RSS 阅读器，速度快，还带一套面向 AI Agent 的 CLI，让 Agent 能在终端里读取、搜索和筛选你的订阅源。
A fast, native RSS reader — plus an agent-facing CLI so your AI agent can read, search and triage your feeds from the shell.
v0.16.1发布于 2026-09-23 01:53 UTC588star67fork同步于 2026-11-07 07:33 UTC
brew install --cask l0ng-ai/papr/papr
brew install l0ng-ai/papr/papr-cli 硬件要求按 README 说明
README 未说明硬件要求，仅说明提供 macOS、Windows、Linux 桌面安装包。
本地优先的原生 RSS 阅读器，附带面向 AI Agent 的 CLI，让 Agent 在 shell 里读写和整理订阅
用 Rust 写的原生 RSS 阅读器，速度快，还带一套面向 AI Agent 的 CLI，让 Agent 能在终端里读取、搜索和筛选你的订阅源。
核心能力
- 桌面端与 CLI 共用 papr-core 和同一本地数据库
- CLI 输出 TOON 格式，比 JSON 省约 40% token
- 支持全文抓取、搜索、标记已读/星标/稍后读
- 提供 papr-rss skill，Agent 按需加载识别订阅任务
- 支持 FreshRSS / Miniflux 同步与 OPML 导入导出
容易踩的坑
内容由模型根据 README 生成，可能不完整或不准确，以官方文档为准。看原文 README
- v0.16.1Papr v0.16.1最新正式版2026-09-23 · 2026-09-23 01:53 UTC
Fixes
- Fix startup failure when opening databases previously used by development builds with schema versions 20 or 21. Existing feeds, articles, favourites, custom icon flags, and feed positions are preserved.
- Show an explanatory startup dialog when the database cannot be opened, instead of aborting the application.
- Rebuild the complete in-app update manifest after platform builds finish, preventing parallel uploads from dropping macOS or Linux update entries.
If v0.16.0 cannot start, download and install v0.16.1 over the existing application. Do not delete your Papr database.
Full changelog: https://github.com/l0ng-ai/papr/compare/v0.16.0...v0.16.1
- v0.16.0Papr v0.16.02026-09-23 · 2026-09-23 01:09 UTC
查看更新说明
What's new
- Image viewer: drag images and zoom in to inspect details. (#132)
- Native Save As dialogs: choose where to export OPML subscriptions and save article images. Cancelling a dialog no longer reports a successful save. (#133)
- AI connection test: check your API key, model, and Base URL directly in Settings, with response timing and useful error messages. Empty responses and outdated test results are no longer shown as successful connections. (#138)
- AI provider profiles: remember each provider's model and Base URL when switching providers. Configuration changes are saved atomically to prevent mixed settings or lost drafts. (#139)
Fixes
- Bundle reader fonts locally, removing the external font request that could delay startup on restricted or offline networks. (#136)
- Improve Linux AppImage compatibility with newer Wayland/Mesa installations and respect an explicitly configured
GDK_BACKEND. Repacked AppImages are re-signed for in-app updates. (#137)
Downloads
- macOS: choose the
aarch64.dmgfor Apple Silicon orx64.dmgfor Intel. - Windows: use the
.msior-setup.exeinstaller. - Linux: use the
.AppImage,.deb, or.rpmpackage. - CLI: platform-specific
papr-*.tar.gzand Windows.ziparchives are included.
Thanks to @expoli and @PrinOrange for their contributions.
Full changelog: https://github.com/l0ng-ai/papr/compare/v0.15.0...v0.16.0
- v0.15.0Papr v0.15.02026-07-11 · 2026-07-11 11:50 UTC
查看更新说明
Minor release.
Features
- Per-feed and global default article open mode, like Fluent Reader's per-source open target (#112, closes #110). Choose how articles open — RSS text, auto-extracted full text, or the original page in the embedded web view:
- Globally in Settings → Reading, where the old "Auto-extract full text" toggle is now a three-way "Open articles as" select (the previous preference is migrated automatically).
- Per feed in Settings → Subscriptions, with each feed defaulting to the global setting. A feed set to "Full text" always extracts the page's own text; a feed set to "RSS text" opts out of global auto-extraction.
See the assets below to download and install this version.
- Per-feed and global default article open mode, like Fluent Reader's per-source open target (#112, closes #110). Choose how articles open — RSS text, auto-extracted full text, or the original page in the embedded web view:
- v0.14.0Papr v0.14.02026-07-11 · 2026-07-11 06:38 UTC
查看更新说明
Minor release.
Features
- Full-screen image viewer in the reader: click any article image to open it full-screen, cycle through the article's images with the on-screen arrows or ←/→, and close with Esc or a backdrop click (#108).
- New "System" appearance mode that follows the OS light/dark setting (#106).
- Appearance split into palette × mode, with six themes to choose from (#102).
Fixes
- FreshRSS: read state is now reconciled against the whole library instead of only the most recent 1000 items, so unread counts no longer run far above the server's (#105).
- Reader no longer shows a redundant, unconstrained hero image on top of articles that already lead with their own image or video (#104).
- Focus mode preserves the article-list scroll position on exit instead of jumping back to the top (#107).
- Feeds without stable guids are de-duplicated by title + date rather than a volatile URL, preventing repeated articles (#103).
See the assets below to download and install this version.
- v0.13.1Papr v0.13.12026-07-11 · 2026-07-11 03:08 UTC
查看更新说明
Patch release.
Fixes
- Feeds from full-archive sources (e.g. Hugo
index.xml) no longer re-surface already-read articles as unread after the daily retention cleanup runs (#98).
See the assets below to download and install this version.
- Feeds from full-archive sources (e.g. Hugo
用户点评
登录用户的真实使用记录，可按版本和硬件筛选共 0 条
还没有人点评。第一个说出真实体验的人最有价值。
Papr
A fast, native RSS reader — and a CLI your AI agent can actually drive.
Papr is two front-ends over one local database:
- A desktop reader — fast, native, offline-first. No account, no cloud.
papr, an agent-facing CLI — so an autonomous agent (Claude Code, Codex, OpenCode…) can read, search and triage your feeds straight from the shell. Jump to it ↓
Both read the same database through the shared papr-core crate, so the app and
the CLI can never drift apart.
The desktop app
- Feeds & folders — subscribe, organize, and import/export OPML.
- Smart views — All, Unread, Starred, and Read Later, with live counts.
- Tags & rules — color-coded tags and rules that tag new articles automatically.
- Full-text — fetch and clean the complete article when a feed ships only a summary.
- AI — summaries, ask-the-article Q&A, and digests. Bring your own API key.
- Audio — a built-in player that follows you from article to article.
- FreshRSS sync — keep read state in step with a FreshRSS server.
- Local-first — everything stays on your machine. No account, no cloud.
- Localized — English, Japanese, and Simplified Chinese.
Install
| Platform | How |
|---|---|
| macOS | brew install --cask l0ng-ai/papr/papr — or grab the .dmg |
| Windows | Download the .msi installer |
| Linux | Download the .AppImage or .deb |
All packages live on the latest release. The macOS builds are Developer ID signed and notarized.
papr — your feeds, handed to your agent
This is what makes Papr different. papr is a command-line companion built for
autonomous agents to drive over the shell. Point your agent at it and it can
work your feeds with no GUI:
- Read —
feeds,list,read, full-textsearch - Triage —
markread/star/later,extractfull text,refresh - Manage — subscriptions, folders, tags, rules, highlights, OPML
- Sync — FreshRSS / Miniflux
Run bare papr and it prints your unread dashboard plus the next useful
commands, so the agent orients with zero manual. Output is
TOON — ~40% fewer tokens than JSON — with definitive
counts and structured exit codes.
$ papr
unread: 206 starred: 17 later: 0 feeds: 15
articles[10]{id,feed,title,flags,date}:
3664,V2EX,[Java] 使用 kkRepo 搭建 Maven 私服,unread.star,"2026-06-25"
...
help[4]: Run `papr read <id>` to read an article's full text, ...
Hand it to your agent — one line
The bundled papr-rss skill loads on demand
when an agent recognizes a feed-related task, so it costs nothing until you use
it. Install it with skills:
npx skills add https://github.com/l0ng-ai/papr/tree/main/skills/papr-rss
Want the agent proactively aware of your feeds every conversation? papr setup
wires up an ambient SessionStart hook (Claude Code, Codex, OpenCode).
Install the CLI
| Platform | How |
|---|---|
| macOS / Linux | brew install l0ng-ai/papr/papr-cli |
| Windows | Download papr-x86_64-pc-windows-msvc.zip, unzip, drop papr.exe on your PATH |
| Any | Prebuilt papr-<target>.tar.gz from the latest release, or cargo build --release -p papr-cli |
Full command reference, agent setup, and install options → docs/cli.md