hyprwhspr-rs
hyprwhspr-rs is a free, MIT-licensed voice dictation daemon for Linux with Hyprland. The README calls it a "Rust implementation of hyprwhspr" by better-slop. Press a Hyprland bind, speak, and the text pastes into the active window through the compositor. Transcription runs with local whisper.cpp or Parakeet, or Groq, Gemini or an OpenAI-compatible endpoint. One caveat: Hyprland only.
What does hyprwhspr-rs do?
hyprwhspr-rs is described in its README as a "Rust implementation of hyprwhspr", built for Hyprland and Omarchy, with optional Waybar and Walker/Elephant menu integrations. It runs as a background daemon. Hyprland owns shortcut capture through bind and bindr lines in hyprland.conf, and triggers hyprwhspr-rs record start, record stop or record toggle over a local socket. The daemon records microphone audio, runs it through a transcription provider, and pastes the result into the active window.
Pasting is done through Hyprland IPC: the daemon reads activewindow, chooses whether Shift is needed for a hardcoded list of programs, and runs dispatch sendshortcut. Fallbacks are a Wayland virtual keyboard client or a simulated keypress paste if IPC fails. Word overrides and an inverse text normalisation step (from FluidInference's text-processing-rs) rewrite spoken forms into typed symbols and punctuation.
Key facts
- Licence: MIT on the code.
- Platforms: Linux with Hyprland (and Omarchy, which is Hyprland-based). GNU-only binaries, no musl releases are published.
- Transcription providers: whisper.cpp (local), Parakeet TDT via ONNX (local), Groq, Gemini, and any OpenAI-compatible endpoint.
- Requirements: whisper.cpp with
whisper-clionPATH, libudev and pkg-config for hotplug detection. - Install channels:
cargo install hyprwhspr-rsfrom crates.io, Nix (NixOS and nixpkgs), or from source withcargo build --release. - Interactive installer:
hyprwhspr-rs installlays down a systemd unit and the Waybar module, with optional Walker/Elephant integration. - Fast VAD trimming (optional) that reduces inference cost and output latency.
- Daemon control commands:
hyprwhspr-rs record {start|stop|toggle|status}. - An optional ChatGPT path, through Codex auth at
~/.codex/auth.json.
What does it replace, and where does it fall short?
hyprwhspr-rs replaces Wispr Flow for Hyprland users: a shortcut-driven dictation loop that pastes into the focused window. The README's own topics list is "hyprland, rust, whisper".
Where it falls short: it is Hyprland only, by design. The README explains the choice: grabbing a keyboard device with EVIOCGRAB would make the tool the sole recipient of events, so Hyprland is left to own shortcut capture through its own bind lines. That design cannot be lifted onto GNOME, KDE or X11 without a different approach. Only GNU builds are published, no musl. Parakeet is listed as fast but not as accurate as whisper or Gemini by the README's own note.
How does hyprwhspr-rs run?
cargo install hyprwhspr-rs installs the binary. Then hyprwhspr-rs install writes the systemd unit (ExecStart points at the binary you ran install from) and optionally the Waybar module. Hyprland needs a few bind and bindr lines in hyprland.conf to call record start, record stop and record toggle. The daemon is left running in the background.
On NixOS, add services.hyprwhspr-rs.enable = true. The nixpkgs package can be overridden with whisper-cpp and onnxruntime set to cudaSupport = true for an NVIDIA box, or rocmSupport on AMD. API keys for Groq or Gemini go in an environmentFile. From source: git clone, cargo build --release, then copy the binary into /usr/local/bin/.
Who is hyprwhspr-rs for?
Linux users on Hyprland or Omarchy who want a dictation shortcut that respects the compositor's input stack. Someone on GNOME, KDE or X11 should pick Handy or another entry on the open-source alternatives to Wispr Flow list. Someone who wants a desktop app with a tray icon should also look elsewhere; this is a daemon, a systemd unit and a Waybar module.
What limits does the README state?
GNU-only binaries, no musl releases. Hyprland only, by the design the README spells out. Parakeet is marked as fast but less accurate than whisper or Gemini. The installer writes the systemd unit with an absolute ExecStart pointing at the binary you ran install from; a manually copied unit template needs hyprwhspr-rs resolvable by systemd. Audio start and stop sounds can go missing in packaging setups, in which case HYPRWHSPR_ASSETS_DIR points at the installed assets. whisper-cli must be on PATH.
Questions people ask
Is hyprwhspr-rs open source?
Yes. The code is MIT, an OSI approved licence. The project is published on crates.io, with the binary installable through cargo install. The README does not mention any trademark carve-out on the name, and the Nix path packages the binary straight from source.
Can hyprwhspr-rs be used commercially?
Yes. MIT allows commercial use, modification and redistribution, provided the licence travels with the code. If you point the daemon at Groq or Gemini for transcription, that provider's pricing and retention apply. Local whisper.cpp and Parakeet keep the audio on the machine.
Does hyprwhspr-rs work offline?
Yes, with whisper.cpp or Parakeet selected as the provider. The daemon then does no network call for transcription. Switching to Groq, Gemini or an OpenAI-compatible endpoint sends the audio over the network to that endpoint.
How does hyprwhspr-rs compare with other Wispr Flow alternatives?
It is the Hyprland-first option. Handy is a cross-desktop app with installers for macOS, Windows and Linux and a Tauri UI. hyprwhspr-rs is a daemon with a Waybar module and systemd unit, designed around Hyprland shortcut capture. The open-source alternatives to Wispr Flow page sets both options side by side with VoiceInk, SpeakOfLow and OpenWhispr.
Which transcription provider should I pick?
The README lists trade-offs: Groq with Whisper is cheap and fast; Gemini is slower but offers better output formatting; Parakeet TDT is very fast locally but not as accurate as whisper or Gemini. whisper.cpp is the local fallback that works offline. The choice depends on whether network calls are acceptable and which failure mode is cheaper.
Sources
- hyprwhspr-rs README and LICENSE: github.com/better-slop/hyprwhspr-rs, read 2026-10-26. The README is the source for every feature, provider, install path and limit above, including the design note about Hyprland input capture.
- crates.io listing: crates.io/crates/hyprwhspr-rs, read 2026-10-26.
More voice dictation apps
A free, offline push-to-talk dictation app for macOS, Windows and Linux that runs Whisper or NVIDIA Parakeet on your own machine and pastes the text at your cursor, under MIT.
A GPL-3.0 macOS dictation app that runs Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper on the device, with command and write modes, per-app prompts, and an optional closed local enhancement model.
A dictation and meeting-notes app under MIT that runs Whisper or Parakeet locally or cloud models with your own key, detects Zoom, Teams and FaceTime calls, and labels speakers on device.
A macOS dictation app under GPL-3.0 whose README names Superwhisper and Wispr Flow as the tools it replaces, with local models, per-app modes, a personal dictionary and an AI assistant mode.
A speech-to-text app inside the Epicenter monorepo under AGPL-3.0 that records, transcribes with a provider you choose, cloud or local, optionally polishes the text, and pastes it at the cursor.
An Apache-2.0 macOS dictation app that calls Gemini 3.5 Transcribe with your own key, writes audio to disk from the first millisecond for crash safety, and inserts polished text at your cursor without an account.
Added October 26, 2026. Every claim above comes from the project's README, LICENSE or model card, read on October 26, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/better-slop/hyprwhspr-rs.
Want hyprwhspr-rs running on your own servers?
Reveneau sets it up on your hardware, connects it to your product, and builds the software around it.