How to Turn a PDF into Audio: Text Layers and Scans

Copy text from several pages to check the file

Open the document in a PDF reader. Clear letters on the screen do not necessarily mean the file contains accessible text: the page could be a photograph or scan. Selecting and copying a passage helps you check what can actually be extracted.

Test more than the cover. A single PDF can combine text pages with scanned inserts. Successfully copying one paragraph does not mean every page is prepared the same way.

1. Choose a paragraph near the start of the main text and try selecting individual words.
2. Paste it into a plain text editor and compare it with the original.
3. Repeat the check in the middle and near the end of the document, and on a page with columns if there are any.
4. Check that words, spaces and sentences remain in the correct order.

Use OCR outside TGRadio to prepare scanned pages

OCR recognises letters in images and converts them into text. A scan needs a separate application with this feature. Renaming a file or saving an image as a PDF does not create a text layer.

If you have the original document, try obtaining a text-based PDF or a supported ebook from it. This can avoid the need to recognise a scan. TGRadio also accepts EPUB, FB2 and TXT, but changing the format alone does not correct errors.

1. Open the scan in an OCR application and select the document language.
2. Recognise the pages you need and compare the result with the original, especially names, numbers and punctuation.
3. Save the result as a PDF with a text layer, or prepare a checked TXT file.
4. Copy several paragraphs from the final file again to verify the result.

Fix text that comes out in the wrong order

In PDFs with columns, footnotes and tables, extracted text may follow a different order from the visible layout. For example, a line from the left column might be followed by one from the right. Text-to-speech will not automatically reconstruct the intended meaning.

If copied text contains broken words, repeated headers or page numbers inside sentences, prepare a clean text version. Fix line-break hyphenation, restore paragraph order and remove interruptions that make listening difficult. Compare your edits with the original so you do not lose content.

When errors are widespread, finding another source file may be more practical. Converting the same PDF to another format does not guarantee a fix.

Check the password, file size and upload method

Password-protected PDFs are not supported by TGRadio. Obtain an accessible version without a password from the source. If document settings prevent copying, that alone does not prove the pages are scans.

The upload limit is 10 MB per file. For a larger document, you can prepare a separate part in a PDF editor and check its text again. Reducing image quality does not add a missing text layer.

After signing in, add a book in the book narration section using a file or public link, or send it to the bot. A link must provide an accessible file or an HTML page containing text; login requirements and protected downloads may prevent importing. Uploaded books are private and are not automatically published.

Check the first available audio fragment

Book narration is currently free for all users, with no monthly character limit. Upload your prepared PDF and check the first audio fragment before continuing.

You can start listening as soon as the first audio fragment is ready. Compare it with the source text. If words are missing or phrases are out of order, check the document first. Changing the voice does not repair the source and requires new narration. Changing playback speed does not require new narration.

Use documents you have the right to use. During narration, the text of each fragment is sent to a speech synthesis service.