# Omarchy Quattro - voxtype-bin (push-to-talk dictation) # Upstream: https://voxtype.io / https://github.com/peteonrails/voxtype (MIT). # Omarchy integrates Voxtype for the SUPER+SPACE row Install > AI > Dictation # (upstream bin/omarchy-voxtype-install runs `omarchy-pkg-add wtype voxtype-bin`), # so the RPM must be named voxtype-bin. Omarchy's default config uses the Whisper # engine (CPU or Vulkan), so this repacks only the Whisper + OSD prebuilt # binaries, not upstream's ONNX/CUDA/MIGraphX payload. %global debug_package %{nil} # Prebuilt binaries carry upstream build-host RPATHs (dlopened from their own # directory); harmless, so skip the check. %global __brp_check_rpaths %{nil} # Optional accelerators (Vulkan, the GTK4/Quickshell OSD) are dlopened at # runtime; keep them out of auto-Requires so the package installs without them. %global __requires_exclude ^(libvulkan\.so.*|libgtk-4\.so.*|libgtk4-layer-shell.*)$ Name: voxtype-bin Version: 1.1.0 Release: 1%{?dist} Summary: Push-to-talk voice-to-text for Linux (Whisper; prebuilt binaries) License: MIT URL: https://voxtype.io Source0: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/config/default.toml#/%{name}-config-%{version}.toml Source1: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/packaging/systemd/voxtype.service#/%{name}-service-%{version} Source2: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/packaging/completions/voxtype.bash#/%{name}-bash-%{version} Source3: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/packaging/completions/voxtype.zsh#/%{name}-zsh-%{version} Source4: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/packaging/completions/voxtype.fish#/%{name}-fish-%{version} Source5: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/LICENSE#/%{name}-LICENSE-%{version} Source6: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/README.md#/%{name}-README-%{version}.md Source7: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/packaging/voxtype-configure.desktop#/%{name}-configure-%{version}.desktop Source8: https://raw.githubusercontent.com/peteonrails/voxtype/v%{version}/packaging/scripts/voxtype-configure-launcher#/%{name}-configure-launcher-%{version} Source9: https://github.com/peteonrails/voxtype/archive/refs/tags/v%{version}.tar.gz#/%{name}-%{version}.tar.gz Source10: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-baseline#/%{name}-%{version}-baseline Source11: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-avx2#/%{name}-%{version}-avx2 Source12: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-avx512#/%{name}-%{version}-avx512 Source13: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-vulkan#/%{name}-%{version}-vulkan Source14: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-osd#/%{name}-%{version}-osd Source15: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-osd-gtk4#/%{name}-%{version}-osd-gtk4 Source16: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-osd-quickshell#/%{name}-%{version}-osd-quickshell Source17: https://github.com/peteonrails/voxtype/releases/download/v%{version}/voxtype-%{version}-linux-x86_64-audio-bridge#/%{name}-%{version}-audio-bridge ExclusiveArch: x86_64 Provides: voxtype = %{version}-%{release} Requires: alsa-lib Requires: curl Requires: glibc Requires: libgcc Recommends: wtype Recommends: wl-clipboard Recommends: vulkan-loader Recommends: gtk4-layer-shell %description Voxtype is push-to-talk voice-to-text for Linux, transcribed locally with Whisper. This package repacks the upstream prebuilt x86_64 binaries (CPU baseline/AVX2/AVX-512 plus the Vulkan GPU build and the on-screen mic visualizer). It is the dictation backend for Omarchy's Install > AI > Dictation menu row. %prep %setup -q -c -T -n %{name}-%{version} tar -xzf "%{SOURCE9}" cp "%{SOURCE5}" LICENSE %build : %install # Whisper CPU/GPU binaries live under %{_libdir}/voxtype; /usr/bin/voxtype is # chosen per-CPU in %post. install -Dm755 "%{SOURCE10}" %{buildroot}%{_libdir}/voxtype/voxtype-baseline install -Dm755 "%{SOURCE11}" %{buildroot}%{_libdir}/voxtype/voxtype-avx2 install -Dm755 "%{SOURCE12}" %{buildroot}%{_libdir}/voxtype/voxtype-avx512 install -Dm755 "%{SOURCE13}" %{buildroot}%{_libdir}/voxtype/voxtype-vulkan install -d %{buildroot}%{_bindir} # Default user-facing symlink; %post retargets it to the best CPU backend. ln -sf %{_libdir}/voxtype/voxtype-avx2 %{buildroot}%{_bindir}/voxtype # OSD launcher + frontends. The launcher resolves /proc/self/exe and probes its # parent directory, so it finds the frontends without PATH gymnastics. install -Dm755 "%{SOURCE14}" %{buildroot}%{_libdir}/voxtype/voxtype-osd install -Dm755 "%{SOURCE15}" %{buildroot}%{_libdir}/voxtype/voxtype-osd-gtk4 install -Dm755 "%{SOURCE16}" %{buildroot}%{_libdir}/voxtype/voxtype-osd-quickshell install -d %{buildroot}%{_bindir} ln -sf %{_libdir}/voxtype/voxtype-osd %{buildroot}%{_bindir}/voxtype-osd install -Dm755 "%{SOURCE17}" %{buildroot}%{_bindir}/voxtype-audio-bridge # Quickshell QML tree (frontend looks under /usr/share/voxtype/quickshell) and # the example OSD styles/recipes, copied whole from the source archive. install -d %{buildroot}%{_datadir}/voxtype cp -a voxtype-%{version}/quickshell %{buildroot}%{_datadir}/voxtype/ find %{buildroot}%{_datadir}/voxtype/quickshell -type f -exec chmod 644 {} + find %{buildroot}%{_datadir}/voxtype/quickshell -type d -exec chmod 755 {} + install -d %{buildroot}%{_datadir}/voxtype/osd %{buildroot}%{_datadir}/voxtype/osd-recipes cp -a voxtype-%{version}/examples/osd-packages/. %{buildroot}%{_datadir}/voxtype/osd/ cp -a voxtype-%{version}/examples/osd-recipes/. %{buildroot}%{_datadir}/voxtype/osd-recipes/ find %{buildroot}%{_datadir}/voxtype/osd -type f -exec chmod 644 {} + find %{buildroot}%{_datadir}/voxtype/osd-recipes -type f -exec chmod 644 {} + install -Dm755 "%{SOURCE8}" %{buildroot}%{_bindir}/voxtype-configure-launcher install -Dm644 "%{SOURCE7}" %{buildroot}%{_datadir}/applications/voxtype-configure.desktop install -Dm644 "%{SOURCE0}" %{buildroot}%{_sysconfdir}/voxtype/config.toml install -Dm644 "%{SOURCE1}" %{buildroot}%{_prefix}/lib/systemd/user/voxtype.service install -Dm644 "%{SOURCE6}" %{buildroot}%{_datadir}/doc/%{name}/README.md install -Dm644 "%{SOURCE2}" %{buildroot}%{_datadir}/bash-completion/completions/voxtype install -Dm644 "%{SOURCE3}" %{buildroot}%{_datadir}/zsh/site-functions/_voxtype install -Dm644 "%{SOURCE4}" %{buildroot}%{_datadir}/fish/vendor_completions.d/voxtype.fish %post # Pick the best CPU backend for /usr/bin/voxtype, but preserve a GPU backend the # user selected with `voxtype setup gpu --enable`. if [ -f /usr/bin/voxtype ]; then target=$(readlink -f /usr/bin/voxtype 2>/dev/null || echo "") case "$target" in */voxtype-vulkan) ;; *) if grep -q avx512f /proc/cpuinfo 2>/dev/null; then ln -sf %{_libdir}/voxtype/voxtype-avx512 /usr/bin/voxtype elif grep -q avx2 /proc/cpuinfo 2>/dev/null; then ln -sf %{_libdir}/voxtype/voxtype-avx2 /usr/bin/voxtype else ln -sf %{_libdir}/voxtype/voxtype-baseline /usr/bin/voxtype fi ;; esac fi %files %license LICENSE %config(noreplace) %{_sysconfdir}/voxtype/config.toml %{_bindir}/voxtype %{_bindir}/voxtype-osd %{_bindir}/voxtype-audio-bridge %{_bindir}/voxtype-configure-launcher %{_libdir}/voxtype/ %{_datadir}/voxtype/ %{_datadir}/applications/voxtype-configure.desktop %{_prefix}/lib/systemd/user/voxtype.service %{_datadir}/doc/%{name}/README.md %{_datadir}/bash-completion/completions/voxtype %{_datadir}/zsh/site-functions/_voxtype %{_datadir}/fish/vendor_completions.d/voxtype.fish %changelog * Thu Oct 08 2026 whelanh - 1.1.0-1 - Initial Fedora package: voxtype dictation backend for Omarchy (v1.1.0)