Open a text PDF, keep only the part you want to hear, check the order, then listen and download WAV audio. Everything runs in your browser — your PDF is not uploaded.
1 Open a text PDF
No document selected. Up to 10 MB and 30 pages.
2 Keep the text that matters
Settings apply to the selected pages. Nothing is removed automatically. For mixed layouts, prepare and edit the text in smaller groups.
0 characters
3 Make the audio
First use downloads a voice model of about 25 MB plus voice and language files. Depending on your connection the first part can take from a few seconds to about a minute; the download progress is shown below. Later generations reuse the loaded voice. English text only; scanned pages need OCR first.
Up to 1,000 characters per generation (about 1–2 minutes of speech). Longer text is never cut silently — choose a shorter excerpt.
Open your PDF. Choose a PDF with selectable text. It is read locally by your browser.
Choose what to hear. Set where reading starts and stops on the page to skip running headers, footers and page numbers. For two-column papers, pick “Left column, then right”.
Check the narration. Press Prepare text, then fix anything in the text box — remove a citation, spell out an abbreviation, or reorder a paragraph.
Generate and listen. The first part plays within seconds while the rest is still being made. Download each part, or the full WAV when it is done.
Why reading order matters
Many “read PDF aloud” tools read whatever text comes first in the file. On papers, reports and handouts that often means the page header, a footnote or the right column interrupting the left one. Here you see the page, mark the area to read and the column order, and review the exact words before any audio is made — so what you hear is the text you meant.
What it does not do
It does not run OCR on scanned pages, interpret equations, describe images or summarize tables. Voices are English only. Audio is generated in parts of up to 1,000 characters so it stays fast on ordinary laptops; long documents are converted section by section.
Frequently asked questions
Is my PDF uploaded?
No. The PDF is opened and read in your browser, and the speech is generated on your device. Your PDF and narration text are not sent to this website. See the privacy page for the files your browser downloads.
Why does the first audio take longer?
The first time you generate audio, your browser downloads a voice model of about 25 MB plus voice and language files, which can take up to about a minute on a slower connection. After that the files are cached, and the first part is usually ready in a few seconds on a typical laptop.
How much text can I turn into audio at once?
Up to 1,000 characters per generation, roughly one to two minutes of speech. Prepare a page or a section, generate it, then move to the next one. Longer text is never cut silently.
Can it read two-column PDFs in the right order?
Yes. Choose “Left column, then right” to read the left column before the right one, and set the reading area to skip headers, footers and page numbers. You can also edit the text before generating.
Does it work with scanned PDFs?
Only PDFs with selectable text work. A scanned page is an image, so it needs OCR first. If no text is found in the reading area, the tool tells you instead of producing empty audio.
What audio format do I get?
A WAV file. You can listen to and download each part as soon as it is ready, then download the full WAV when all parts are finished.