Read a PDF Aloud with Python — PyPDF2 and pyttsx3 Offline Text-to-Speech

A short Python script that reads a PDF aloud, page by page: it extracts the text with PyPDF2 and speaks it with pyttsx3, which uses your operating system’s built-in (offline) text-to-speech voices.

Overview

This script extracts text from a PDF file using PyPDF2 and converts it to speech using pyttsx3. It reads each page sequentially, with a brief pause between pages for natural pacing. A preferred voice (female by default) is selected automatically if available.

Features

  • Reads all pages of a PDF aloud using the system’s text-to-speech engine
  • Tries to select a voice whose name or ID contains “female” (see the note on voice selection below)
  • Configurable reading speed (words per minute) and volume
  • Prints progress to the console as each page is read
  • Short pause between pages for a more natural listening experience
  • Error handling for missing files and unexpected issues

Requirements

PyPDF2
pyttsx3

Install dependencies:

pip install PyPDF2 pyttsx3

Usage

Story.pdf is not included in the repository. Place your own PDF named Story.pdf in the folder you run the script from (or update the path), then run:

python StoryReader.py

The default configuration reads Story.pdf with a female voice at speed 150 and volume 0.9.

Changing the PDF File

Update the file path in the __main__ block:

read_pdf_aloud("path/to/your/file.pdf", voice_id=chosen_voice, rate=150, volume=0.9)

Parameters

Parameter Default Description
file_path "Story.pdf" Path to the PDF file
voice_id Female voice (auto-selected) System voice ID
rate 150 Speaking speed in words per minute
volume 0.9 Volume level (0.0 to 1.0)

File Structure

StoryReader/
└── Console_Code/
    └── StoryReader.py    # Main script

Notes

  • pyttsx3 uses the operating system’s built-in TTS engine (SAPI5 on Windows, NSSpeechSynthesizer on macOS, eSpeak on Linux).
  • Pages with no extractable text are silently skipped.
  • Press Ctrl+C to interrupt playback.
  • PyPDF2 is no longer developed under that name; its successor is pypdf, which provides the same PdfReader class.

Voice selection

select_voice looks for the word “female” in each installed voice’s name or ID. Standard Windows voice names such as “Microsoft Zira Desktop” do not contain that word, so the function usually falls back to voices[1], the second installed voice. On a default English (US) Windows installation, that is typically Zira. If only one voice is installed, voices[1] raises an IndexError; change the fallback to voices[0] or pass a specific voice ID.

Read all Python Projects and Code Examples guides online · Project overview · All developer projects