Search, read, and listen to your documents offline

Currently, offline models are used for audiobook generation. Reading, import, organization, and search are local document tools.

Reading and search work offline. Network access is used when you choose to download a voice model, download a release, or open an external resource. If you do not use audiobook generation, voice model downloads are not needed.

English voices use Kokoro because it currently offers strong quality among fully offline voice models. The goal is natural-sounding audio without cloud processing, subscriptions, API keys, or per-character billing.

Arabic is supported with separate offline models. Piper is the smaller cross-platform option, while SILMA is available on Linux x64 desktop through an optional runtime download.

Papercut does not claim ownership of generated audio. You are responsible for rights in the source text, any reference voice or recording used, and how exported audio is shared. Papercut saves audiobook output locally and does not upload it to a service.

Yes. Saved audiobooks can be exported as Papercut audiobook bundles, then shared, transferred to another device, backed up, or imported back into Papercut later. If you share, publish, sell, or broadcast generated audio, label it as AI-generated and do not use it to impersonate or mislead.

No. Reference-voice features should only be used with voices and recordings you have rights or permission to use.

Papercut imports HTML, EPUB, and PDF files. PDF search, Find, and audiobook generation require embedded text; image-only PDFs can be viewed when supported but are not searchable because Papercut does not include OCR yet.

Papercut is available in English, Arabic, Simplified Chinese, Spanish, French, Hindi, Italian, and Brazilian Portuguese.

Kokoro provides voices for English, Mandarin Chinese, Spanish, French, Hindi, Italian, and Brazilian Portuguese. Arabic uses separate offline Piper and Supertonic models, with SILMA available on Linux x64 through an optional runtime.

Android: Android 8.0 / API 26+ on arm64-v8a devices.

iOS: planned, App Store release coming soon.

macOS: macOS 10.13+ on Apple Silicon or Intel Macs.

Linux: amd64 Linux with WebKitGTK and GStreamer available; release artifacts are built on Ubuntu 24.04.

Windows: Windows 10+ x64 with the Microsoft Edge WebView2 runtime.

Enable Install unknown apps in Settings for your browser or file manager, then open the APK. Reading and search work offline. Network access is only used if you choose to download a voice model; if you do not use audiobook generation, network access is not needed.

Requires amd64 Linux with WebKitGTK and GStreamer available. Release artifacts are built on Ubuntu 24.04.

Debian-based (Ubuntu, Mint): sudo apt install ./Papercut_*.deb

RPM-based (Fedora, openSUSE): sudo rpm -i Papercut-*.rpm

AppImage: make executable (chmod +x) and run directly. On Arch-based systems the AppImage may need WEBKIT_DISABLE_COMPOSITING_MODE=1; see the README for details.

Requires macOS 10.13+ on Apple Silicon or Intel Macs. Open the .dmg and drag Papercut to your Applications folder. Official release builds are Developer ID signed and notarized. If you are testing an unsigned CI or local development artifact, right-click (or Control-click) the app icon and choose Open on first launch.

Pick the build that matches your Mac: Apple Silicon (_aarch64.dmg) for M-series chips, or Intel (_x64.dmg) for Intel-based Macs.

Requires Windows 10+ on x64 hardware and the Microsoft Edge WebView2 runtime. The MSI installer handles setup automatically. If you prefer the portable EXE, run it directly. No admin rights needed.

Start with the README for setup help and platform details. See what's new in the release notes, browse all releases, or contribute on GitHub.