Skip to content
Athena mark Athena mark Athena

Athena Whisper

Athena Whisper is an open-source desktop dictation widget for turning speech into text and inserting it into the currently focused app. It is built for the Athena Home AI workspace and uses faster-whisper for local speech recognition, with platform-native keyboard and clipboard injection for system-wide text input on Linux and Windows.

The goal is a local-first alternative to cloud dictation tools such as Wispr Flow: click a small always-on-top widget, speak naturally, transcribe with Whisper, clean the text, and type it into Codex, Claude Code, opencode, terminals, browsers, chat apps, documents, and other text fields.

What it does

Section titled “What it does”
  1. Focus a text box in any app.
  2. Launch athena-dictate widget.
  3. Click the widget to start recording.
  4. Speak into the microphone.
  5. Click stop.
  6. The app transcribes speech locally with faster-whisper.
  7. The cleaned text is inserted into the previously focused app.

This is designed for hands-free or low-friction text entry in coding agents, shells, browsers, notes, chat, email, and desktop applications.

Features

Section titled “Features”
  • Floating always-on-top dictation widget (PyQt6, with PySide6 fallback if installed)
  • Local Whisper transcription via faster-whisper — CPU-friendly multilingual defaults (base model, int8 quantization)
  • Microphone recording with sounddevice and soundfile (16 kHz mono)
  • Basic dictation cleanup: whitespace normalization and spoken punctuation such as “comma”, “period”, and “new line”
  • Insertion backends for Linux and Windows:
    • X11: clipboard paste, terminal paste fallbacks, direct keystroke typing
    • Windows: clipboard paste via keybd_event, unicode keystroke injection via SendInput
    • Wayland: wl-copy + wtype/ydotool
  • CLI commands for diagnostics, file transcription, one-shot dictation, and insertion testing
  • Configurable defaults through athena-dictate.toml

Privacy posture

Section titled “Privacy posture”

Transcription is local by default — no cloud transcription is required by the current implementation. The current repository implements the faster-whisper backend; it does not include a hosted transcription backend.

Current limitations

Section titled “Current limitations”
  • Linux (X11/Wayland) and Windows are supported; macOS is not yet implemented.
  • Wayland support depends on compositor-specific tools such as wl-copy, wtype, or ydotool.
  • faster-whisper on CPU is practical for short dictation but is not instant large-model streaming ASR.
  • Cleanup is rule-based today; LLM polishing and command mode are future work.
  • System-wide insertion is inherently fragile because every terminal, compositor, and app handles synthetic input differently.

Roadmap

Section titled “Roadmap”
  • Global push-to-talk hotkey
  • Better Wayland support
  • Optional LLM cleanup/polish pass
  • Command mode for editing selected text by voice
  • Personal dictionary and phrase correction
  • Local transcript history
  • Latency benchmarks across tiny, base, and small