avi/S.H.O.N.A.R._Desktop_Companion

avi 7234a62eb5 Debian support: verified build/packaging path + friend install guide

- setup-standalone.sh: falls back to system python3 -m venv + pip when
  uv isn't installed (Debian/Ubuntu path; verified on clean bookworm).
- dist/Dockerfile.debian: debian:bookworm build env for :app:packageDeb
  (needs fakeroot + /usr/share/desktop-directories — both baked in; the
  dir gap made jpackage's postinst fail with exit 3 on minimal systems).
- dist/make-release.sh: clean source tarball from HEAD.
- INSTALL-debian.md: end-to-end install for friends (source tree must
  stay on disk — the app spawns the engine from it via SHONAR_REPO).

Verified on real Debian 12 in Docker: :app:packageDeb builds
shonar-desktop_0.1.0-1_amd64.deb, apt installs it, setup-standalone.sh
provisions the engine (venv+pip path, model fetch, boot smoke test),
and the packaged binary launches under Xvfb.

2026-09-15 07:10:27 -05:00

2 KiB

Raw Permalink Blame History

SHONAR Desktop — installing on Debian (12/13) or Ubuntu

Everything runs locally on your machine. No accounts, no cloud, no telemetry. Audio, transcripts, and settings never leave your computer.

There are two pieces: the desktop app (browse, transcribe, search) and the transcription engine (faster-whisper, bundled model — works offline after setup).

1. Put the source tree where the app can find it

The app auto-starts the bundled engine from the source folder, so keep the unpacked tarball around. Either:

mkdir -p ~/Projects && tar xzf shonar-desktop-*.tar.gz -C ~/Projects \
  && mv ~/Projects/shonar-desktop-* ~/Projects/"Shonar Desktop"

or extract anywhere and export SHONAR_REPO=/that/path (add it to ~/.profile, or to the desktop launcher's Environment= line).

2. Install the app

Option A — .deb package (if you got shonar-desktop_*.deb):

sudo apt install ./shonar-desktop_0.1.0_amd64.deb
shonar-desktop

Option B — run straight from the source tree (needs JDK 17, step 3):

cd ~/Projects/"Shonar Desktop" && ./launch-desktop.sh

3. One-time setup

sudo apt install -y openjdk-17-jdk python3 python3-venv mpv
cd ~/Projects/"Shonar Desktop" && ./setup-standalone.sh

This creates the engine's Python environment, vendors the whisper "base" model into models/hf-cache (one-time download, ~142 MB), and boots the engine once to prove it works.

4. Run it

Launch shonar-desktop (menu entry) or ./launch-desktop.sh. The app starts the engine itself on first launch (takes a few seconds). Point "Choose folder…" at the folder holding your recordings and go.

Notes

mpv is used for click-to-play from search results.
Summaries: if you run Ollama locally the app uses it automatically; without it, transcription still works and summaries are simply skipped.
Engine logs: ~/.config/shonar-desktop/engine.log.
Wayland/Hyprland works; on X11 set nothing special.
Ctrl+= / Ctrl+- zooms the whole app; Ctrl+0 resets.