FonadaLabs, the speech AI division of Fonada has secured No. 1 position for Hindi on the Hugging Face Open ASR Leaderboard rank for its Kapture V3 model.
Kapture V3 recorded an average Word Error Rate (WER) of 6.54% on the Hindi benchmark, placing it ahead of the other leading global models. The result reflects Fonada consistent efforts to develop accurate, cost-efficient speech recognition technology for Indian languages and conversational applications. The achievement is noteworthy in a global AI ecosystem where technology companies such as OpenAI, Google, Microsoft, NVIDIA and IBM are advancing speech and language AI capabilities.
Automatic Speech Recognition (ASR) technology converts spoken language into text and is increasingly being used in customer service, contact centers, transcription, and enterprise automation.
Alongside recognition accuracy, Fonada is focusing on making speech AI accessible and economical for developers and businesses. Kapture V3’s inference cost is approximately ₹0.10 per minute of audio, equivalent to ₹6 per hour, based on the company’s stated pricing. This positions the model as a cost-conscious option in a market where global technology companies such as OpenAI, Google, Microsoft and AWS offer speech recognition and AI services through their respective platforms. While pricing structures vary depending on usage, features and deployment models, Fonada aims to lower the cost barrier for businesses adopting speech AI, particularly for Indian-language and conversational applications.
The model is available through an API and supports 10 Indian languages alongside English, with transcripts in native scripts and segment-level timestamps. These capabilities can help developers build applications for multilingual customer interactions, transcription, and voice-enabled services.
Kapture V3 is part of Fonada broader speech AI roadmap, which includes text-to-speech (TTS), voice cloning, noise cancellation and AI dubbing. The company plans to continue research and development in natural speech generation, Indian-language pronunciation and efficient inference.
Commenting on the achievement, Nitish Gopalani, Co-Founder and CEO of Fonada, said “Reaching the top position on the Hugging Face Open ASR Leaderboard is an important milestone for FonadaLabs and, more importantly, for the Indian speech AI ecosystem. Our focus has always been to build speech technology that understands Indian languages accurately while remaining practical and affordable for real-world enterprise use”
Fonada recognised the support of the Ministry of Electronics and Information Technology (MeitY) and the IndiaAI Mission. According to the company, a 40% compute subsidy supported its model training and experimentation. The company expressed appreciation for the leadership of Shri Ashwini Vaishnaw and Shri S. Krishnan, as well as the contributions of Shri Saurabh Vijay, Shri Sudeep Shrivastava and the IndiaAI Mission team towards improving access to AI computing resources.
The company also acknowledged the role of Hugging Face leaderboard team for enabling transparent model comparisons.
About Fonada
Fonada (Shivtel Communications Pvt. Ltd.) provides business communications infrastructure, omnichannel AI agents and speech AI solutions. Its offerings include cloud telephony, IVR, cloud contact centres, messaging, AI-powered conversational agents, speech-to-text, text-to-speech and other speech processing technologies.
About Hugging Face
Hugging Face is an AI and machine learning platform that enables developers and researchers to share, discover, evaluate and use AI models and datasets. Its Open ASR Leaderboard allows speech recognition models to be compared across multiple datasets and languages, providing a common reference for evaluating their performance.
