pyttsx3

pyttsx3 is a text-to-speech conversion library for Python 3 that, unlike cloud-based alternatives, works entirely offline with no internet connection or per-request latency. It provides a single, intuitive API on top of the native speech engines already present on the host operating system, so the same code speaks text on Windows, macOS, and Linux.

Under the hood it wraps SAPI5 on Windows, NSSpeechSynthesizer and AVSpeech on macOS, and eSpeak/espeak-ng on Linux, letting you choose among installed voices and control speech rate and volume. Speech can be played aloud or saved directly to an audio file, making it a common building block for accessibility tools, assistants, and notification systems.

What You Get

• A single Python API that speaks text using the OS’s native voices
• Fully offline operation with no internet connection or API keys required
• Cross-platform drivers for SAPI5 (Windows), NSSpeech/AVSpeech (macOS), and eSpeak (Linux)
• Control over voice selection, speech rate, and volume
• The ability to save synthesized speech to an audio file

Common Use Cases

• Adding spoken output to desktop assistants and accessibility tools
• Generating voice notifications or alerts without a network dependency
• Rendering text to audio files for offline playback
• Prototyping voice interfaces that must work air-gapped or offline

Under The Hood

Architecture - pyttsx3 exposes a top-level init() factory (pyttsx3/init.py) that returns an Engine (engine.py) which delegates to a platform driver loaded through a driver abstraction (driver.py) with concrete backends in the drivers/ package (sapi5, nsss, avspeech, espeak), plus a Voice value object (voice.py). This driver pattern isolates OS-specific speech APIs behind a uniform engine interface. Tech Stack - Pure Python 3 (3.10+ per project metadata) using setuptools packaging; platform integration relies on pywin32/comtypes for SAPI5 on Windows, pyobjc for the macOS synthesizers, and the system espeak-ng/libespeak library on Linux, so most heavy lifting is delegated to native OS engines. Code Quality - The repository includes a tests/ suite (test_pyttsx3.py, test_engines.py), documentation hosted on Read the Docs, a makefile, and an active release history spanning many versions, reflecting a mature and maintained project. API Design - The API is deliberately minimal and discoverable: init(), say(), runAndWait(), and getProperty/setProperty for rate, volume, and voices, with a one-line pyttsx3.speak() convenience, keeping the barrier to first sound extremely low.