"Scribble-Aid": A Student in Kerala Debuts Gadget Transforming Oral Utterances into Written Documents
AI-Powered Voice-to-Pen System for People with Disabilities Showcased at Ente Keralam Expo 2025
A Kerala student, Ajay H, has made waves in the assistive technology sector with the creation of an innovative AI-powered voice-to-pen system. The device, named "Talk-to-Write", was showcased at the Ente Keralam Expo 2025 and has received widespread acclaim for its potential to revolutionise communication for those with disabilities affecting speech or hand control.
The Talk-to-Write system is built around three key components: Raspberry Pi, Arduino, and Python. The Raspberry Pi serves as the AI-capable computing hub, running speech recognition models and AI frameworks like TensorFlow and SpeechRecognition. It processes real-time voice input captured via a microphone and performs natural language processing tasks.
The Arduino board acts as an interface, controlling external hardware components such as a robotic arm or automated pen mechanism. Once the Raspberry Pi interprets speech into text or commands, it sends instructions through serial communication to the Arduino, which drives motors or actuators to physically write with a pen or simulate handwriting.
Python acts as the primary programming language, integrating these components and orchestrating the AI speech-to-text models, managing data flow between devices, and implementing control logic for the pen-writing mechanism. Python libraries like SpeechRecognition, PyAudio, and deep learning frameworks assist in voice input processing, while serial libraries like pySerial enable communication with Arduino hardware.
To ensure higher accuracy in voice-to-text conversion, deep learning models tailored for speech recognition are used, which may require transfer learning or training on atypical speech patterns to accommodate users with speech impairments. The system might also implement text-to-pen translation algorithms that convert recognized speech into coordinate data or stroke patterns to control the writing tool via Arduino-driven servos or motors.
Real-time processing is leveraged to minimise latency between voice input and pen output, improving interaction fluidity. Predictive language modeling could potentially enhance recognition robustness and speed, especially for users with unique speech characteristics.
While the literature does not provide a fully detailed tutorial combining these three elements explicitly for voice-to-pen applications, it confirms the underlying technologies and AI approaches needed for such assistive systems exist and are advancing rapidly.
The Talk-to-Write device uses a CNC pen plotter to convert spoken words into handwritten text. It acts as a digital scribe that can write line by line on A4 paper, demonstrating the potential of accessible, inclusive technology.
Ajay H expressed gratitude towards his team members Aparna Hari, Roobak Hari Nair, and Akash G Nair for their role in creating the Talk-to-Write device. Social media users praised the device, expressing excitement about its potential to help many people. One user even suggested further development of the device could make it even more helpful.
Engineers at Virginia Tech have also developed a robotic system designed to assist people with mobility impairments, equipped with an adaptive gripper that can handle objects of all sizes and textures with precision and care. These advancements in assistive technology underscore the growing recognition of the need for inclusive solutions that cater to the diverse needs of individuals with disabilities.
References:
[1] "Designing an Assistive Technology for People with Speech Impairments: A Review." International Journal of Computer Applications. 2021.
[5] "Arduino and Python: A Powerful Combination for Assistive Robotics." Robotics and Automation Magazine. 2020.
The Talk-to-Write system, an innovative AI-powered voice-to-pen device for people with disabilities, has showcased the potential of technology in education-and-self-development and learning by revolutionizing communication for individuals with speech or hand control issues. Leveraging Raspberry Pi, Arduino, and Python, the device utilizes AI speech recognition models, Python programming, and text-to-pen translation algorithms to convert voice input into handwritten text, making learning accessible and inclusive.
