A text-to-speech app is an app that reads any text aloud in a natural-sounding voice: a web page, an email, a PDF, an ebook or a photo of a printed page. It is also called a TTS app or a read-aloud app. People with dyslexia, a condition that makes reading hard, use it to follow text that would take them much longer to read. So do people with weak eyesight, and anyone who would rather listen, such as a person commuting by train.
How does it work? First, the app gets the text. From a web page or an ebook, it takes the text directly. From a scanned PDF or a photo of a book page, it uses optical character recognition (OCR), software that recognises the letters in an image and turns them into text. Next, the app prepares the text. It skips page numbers and footnote markers, and decides how to say things like "₹2.5 lakh" or "Dr.". Then a speech AI model turns the text into speech. Older systems joined recorded sounds together, and they sounded robotic. Today's models are trained on many hours of human speech, so they pause at commas and change their tone at a question mark.
Take, for example, a first-year law student in Chandigarh who has dyslexia. Her reading for one week is three judgments, about 120 pages in all. Reading them on her own would take her most of the weekend. She opens the PDFs in the app and listens at 1.3 times normal speed. The app highlights each word on the screen as it reads, and seeing and hearing the words together helps her keep her place. She finishes the reading in about four hours.
Most text-to-speech apps are freemium. The free version reads with standard voices. A paid subscription adds more natural voices, faster speeds and the scanning of printed pages. Phones have a basic read-aloud feature built in, so paid apps compete on voice quality.
Models now also read Hindi, Tamil, Bengali and other Indian languages, as part of the wider language and writing market.
How is it different from speech recognition? Speech recognition turns speech into text. Text-to-speech does the opposite: it turns text into speech. And how is it different from an audiobook? An audiobook is a recording of a person reading one book aloud. A text-to-speech app reads any text that the user gives it, immediately.