The vernacular internet is the part of the internet that Indians use in their own languages, such as Hindi, Tamil, Telugu, Bengali and Marathi, rather than in English. "Vernacular" means the language people use at home. Most of the hundreds of millions of Indians who came online in the last ten years came online this way. They did so on low-cost smartphones, after mobile data became very cheap from about 2016. Many of them read and write mainly in their own language. Apps that were built in English first did not reach them. So Indian companies built apps for these users, in their languages.
Building for the vernacular internet takes more than translating an app's menus. Here is how such an app usually works. When it first opens, it asks the user to choose a language, and then shows every button and screen in that script. Each Indian script is stored in Unicode, the international standard that gives every letter of every script its own number, so that any phone can show it. Many users do not type in their own script at all. They type Hindi in English letters, such as "kaise ho", and the app converts it into Devanagari. This is called transliteration. Many others prefer to speak, so the app accepts voice through speech recognition, which turns speech into text. The content is made in the language too: posts by other users, stories by local writers, and films and shows dubbed for that audience. The app is kept small and quick to load, because many users have phones with little storage and weak networks. It earns mostly from ads and small payments, because these users pay less for subscriptions than English-speaking city users do.
AI models now translate and voice content from one Indian language into another. The government's Bhashini platform, part of India's digital public infrastructure, offers such translation and speech models to any app. Models trained on Indian languages are also a goal of sovereign AI projects.
Take, for example, a vernacular social media app. A shopkeeper in a district town posts a morning greeting in Marathi, and it is shown to other people who read Marathi.
Consider an audio series app. A listener on a bus hears a thriller in Bengali, five minutes at a time, without having to read anything.
For instance, a story writing platform lets a retired bank employee write a novel in Tamil, chapter by chapter, and earn when readers pay to unlock the later chapters.
How is the vernacular internet different from localisation? Localisation means translating an existing app and its content into another language. The vernacular internet is about users, content and habits that start in the language itself: what people post, how they type, and what they pay for. An app that is only translated often fails, because its content and its design are still made for English-speaking city users.
Appears in
- Micro-drama app↗ Video streamingA micro-drama app releases its short episodes in Hindi, Tamil, Telugu and other languages for viewers in smaller towns.
- AI dubbing and lip sync↗ Film production and VFXAI dubbing turns one Hindi film into Tamil, Telugu and Bengali versions without reshooting any scene.
- Audio series and audio drama app↗ Audio series and audiobooksAn audio series app offers its stories in Hindi, Tamil and Bengali for listeners who prefer their own language.
- Audiobook subscription↗ Audio series and audiobooksAn audiobook subscription in India records the same book in several Indian languages, each read by a different narrator.
- Online story writing platform↗ Audio series and audiobooksAn online story writing platform hosts stories written by readers in Marathi, Bengali, Hindi and other Indian languages.
- Vernacular social media app↗ Social mediaA vernacular social media app shows each user posts in the language they chose when they signed up.
- Short-video app↗ Social mediaA short-video app lets users make and watch clips in their own language, and shows them more clips in that language.
- Matrimonial website↗ Social mediaA matrimonial website runs separate sites for each language and community, such as Tamil, Telugu or Marathi users.