An X post about character limits on Sarvam AI’s text-to-speech tool has prompted CEO Pratyush Kumar to offer lifetime access to a user with a speech impairment while promising to step up its broader accessibility efforts. Addressing the feedback publicly, Kumar claimed that cost constraints should never hinder disabled individuals from using AI to navigate daily routines.
It is our lost voice restored. “I have spent years communicating through written notes, often unread or unanswered. At cancer hospitals, I see countless patients struggling to express their suffering. Our pain is ours, but our voice has become someone else’s. For us, Text-to-Speech AI is not a convenience. He added that Sarvam AI’s current per-character pricing model makes regular use expensive for many. Raghunath urged the company to introduce an affordable, unlimited monthly subscription tailored for people with disabilities (Divyangjan). “We are not asking for charity. We are asking for the dignity of speaking for ourselves. Kumar started his tweet by thanking Raghunath, adding, “It is at the heart of what we are attempting at @SarvamAI. He added, “We agree that you shouldn’t have to count characters to be empowered by AI. We have credited your account with a lifetime access to our models. Further, he invited anyone using Sarvam for accessibility to write to the company, promising that the team will work directly with them. It is a technology that uses artificial intelligence and an audio system to convert written text into spoken language. text-to-speech (TTS) works in the exact opposite direction: taking text input and generating audio output While Speech-to-Text (STT) transcribes human speech into written text. Sarvam AI claims it’s text-to-speech technology: “Turn text into speech that feels human, carries emotion, and sounds natural in every interaction. “Hindi, Bengali, Tamil, Telugu, Gujarati, Kannada, Malayalam, Marathi, Punjabi, Odia, and English (Indian accent). Each language supports multiple speaker voices with different characteristics. For many non-speaking individuals, typing out words on a phone is fast, but hearing people do not always stop to read a screen. Modern text-to-speech apps solve this by creating a real-time, two-way bridge. A non-speaking person types their message into an app, which reads it aloud. When the other person speaks back, the app, using speech-to-text technology, converts their speech into text on the screen. Research on voice accessibility platforms published in the Journal of Medical Internet Research (JMIR) shows that real-time transcription interfaces significantly increase daily autonomy and independence for individuals with communication barriers.
In a tweet, a 70-year-old oral cancer survivor appealed to Sarvam AI for more accessible text-to-speech technology, describing it as essential for restoring independence to people with speech disabilities. Raghunath AS said that he lost his ability to speak after fighting oral cancer in 2015. He continued, “India has an estimated 1.99 million people with speech disabilities. A doctor may not read my 100-word note, but could listen to it in under a minute. The tool, powered by its Bulbul V3 model, supports 11 Indian languages, according to its official website. As for the pricing, the base plan starts at ₹ 20 for 10,000 characters.
He also said that the company will eventually launch a formal program regarding the matter. It is also called speech synthesis, voice generation, or “read aloud” technology.
For those who communicate through sign language or gestures, it can be hard to interact with someone who does not know sign language.
Real-time gesture-to-text systems are another way that helps bridge the gap between non-speaking users and others, according to research published via the IEEE Xplore Digital Library.
