Smart gloves powered by AI and technology may soon dissolve the barriers between the deaf and the hearing world.
The idea to work for the deaf community began in Aishwarya Karnataki’s school days, when she met a boy from the special needs section who used sign language. This prompted her to learn it so she could communicate with him. Years later, while studying electronics and building a robot, she thought of creating gloves that could translate sign language into speech. That idea evolved into Glovatrix, a company she started with Parikshit Sohoni in Pune to help the deaf communicate. They began work in January 2020 and launched the company in April 2021.
Together, they developed AI-based gloves that convert sign language into voice so everyone can understand what deaf people want to say. The gloves also have a microphone that converts spoken words into text, helping deaf people understand others. When asked about the name of their startup, Aishwarya said, “We grew up watching a cartoon called Swat Kats. It had a special glove that could shoot missiles. That glove gave us the idea to make our own gloves—not for fighting, but to help people talk to each other.”
AI is key to how the gloves work. They have 12 sensors, each sending six numbers constantly, so the AI receives 72 data points every moment. These numbers form unique patterns for each sign, like “hello” or “how are you?” The AI reads these patterns to identify the sign. Since sign language follows a different word order—like ‘cat mouse ran’ instead of ‘a cat ran after the mouse’—the AI uses natural language processing to translate signs into proper English sentences. It also learns to recognise small differences in how people sign, just as handwriting varies, making communication easier for all.
Regarding the electronics involved, Aishwarya explains, “We use two types of PCBs: a main one on the wrist and 10 custom ones on the fingertips. The main PCB includes a motion sensor, ESP32 chip, microphone, speaker, haptic motor, screen, and supporting components. Each fingertip PCB has a motion sensor, with one fingertip featuring a button. You press the button, perform sign gestures, then press it again. The gestures between presses are sent to the cloud for processing, helping the system distinguish sign language from other hand movements such as scratching.”
The gloves currently support Indian sign language, with plans to add American and British sign languages. The spoken output can be in Hindi, Marathi, English, or other languages, based on user preference. An application on the phone allows users to easily change the settings.
Discussing design challenges, Aishwarya says, “The biggest challenge in developing the app was retrieving data from the hardware—especially from two gloves simultaneously. Initially, we built separate apps for Android and iOS, but our partner, Coreco Technologies, recommended using Flutter for cross-platform development. They even created a basic Bluetooth-connected prototype for free, then later built the full app. Now, the right glove sends its data to the left glove, which aggregates both data sets and transmits them to the phone.”
The PCBs are made in China and arrive with some surface-mount components (SMDs) pre-attached. Remaining parts, sourced from Japan, the US, Taiwan, China, and other countries, are hand-soldered in India. The outer casing is locally 3D printed, and gloves are stitched by a local tailor. All components are assembled and tested in-house.
Speaking about revenue, Aishwarya adds, “We have not sold any device yet. However, we did get support through a CSR initiative from a hotel chain called Royal Orchid Hotels, which has recently started hiring people with hearing and speech disabilities. Through their CSR, they have placed a pre-order for 10 devices as part of a paid pilot project. We will start deploying those next month.”
Looking ahead, the main goal is to enter the market via B2B sales, starting by deploying gloves to organisations. After gaining traction in select locations, the startup will seek investment to grow and expand marketing, sales, and research teams, and build a large sign language gesture database. By the end of 2026, they plan to launch directly to consumers and expand globally by adding American sign language support.
—Nidhi Agarwal
