dictation-toolbox/dragonfly

Speech recognition framework allowing powerful Python-based scripting and extension of Dragon NaturallySpeaking (DNS), Windows Speech Recognition (WSR), Kaldi and CMU Pocket Sphinx

What it solves

Dragonfly is a speech recognition framework for Python that simplifies the creation of custom voice commands. It allows users to integrate speech recognition engines with Python scripts, macros, and applications to automate computer activities, program by voice, or dictate prose.

How it works

Dragonfly treats speech commands and grammar objects as first-class Python objects. It provides a framework for defining rules (such as CompoundRule and MappingRule) that map spoken phrases to specific actions, such as simulating key-strokes or entering text. It can be active based on specific application contexts (e.g., only active when a Notepad window is in focus).

Who it’s for

Developers and power users who want to create complex, voice-controlled automation and programming tools using Python.

Highlights

  • Multi-engine support: Supports Nuance Dragon, Windows Speech Recognition (WSR), Kaldi, and CMU Pocket Sphinx.
  • Cross-platform: Works on Windows, macOS, and Linux (X11).
  • Action framework: Includes built-in actions for text input and key-stroke simulation.
  • Context-aware: Commands can be restricted to specific applications via AppContext.

Related

  • Project
    daanzu/kaldi-active-grammar

    Kaldi‑Active‑Grammar is a Python wrapper for the Kaldi speech‑recognition engine that lets you compile many small grammars and activate only the needed ones per utterance. It ships binary wheels for all major OSes, includes a pre‑trained English model, and provides a backend for the Dragonfly and Caster voice‑control frameworks. Install via pip, download the model, define rules, and dynamically enable them for context‑aware command‑and‑control.

  • Project
    OHF-Voice/speech-to-phrase

    Speech‑to‑Phrase is a local, Home Assistant‑focused speech recogniser that matches spoken input against a set of predefined command templates (including your entity, area and floor names). It builds a finite‑state transducer from language templates, trains a Kaldi language model, and adds a fuzzy post‑processor for robust command detection. The repo holds the last Kaldi‑based version (v1.4.3) supporting 22 languages; active development has moved to a newer rewrite in the OHF‑Voice/apps monorepo.

  • Project
    Uberi/speech_recognition

    A Python library that provides a unified interface for speech recognition, supporting multiple online and offline engines like OpenAI Whisper, Google Speech, and Vosk.

  • Project
    stanfordnlp/dspy

    DSPy is a framework that allows developers to build modular AI systems using Python code instead of manual prompt engineering, automatically optimizing prompts and model weights for better performance.

  • Project
    VOICEVOX/voicevox_core

    The core speech synthesis engine for VOICEVOX, provided as a C and Python library for easy integration into various applications.