India Launches VYOMA Challenge to Drive Multilingual Edge AI Innovation

India has launched the VYOMA Innovation Challenge to accelerate multilingual, voice‑first AI solutions that work offline, aiming to strengthen digital inclusion across diverse languages and regions. The initiative offers prizes worth up to ₹80 lakh and deployment opportunities with government departments.

The challenge has been launched by the Digital India BHASHINI Division (DIBD), under the Digital India Corporation (DIC), Ministry of Electronics and Information Technology (MeitY), in collaboration with Current AI and Kalpa Impact.

The challenge builds on Sunno Sutra, a multilingual, voice-first, open-source handheld AI reference device jointly developed by BHASHINI and Current AI and unveiled at the IndiaAI Impact Summit 2026. Designed as a reference platform, Sunno Sutra combines multilingual language technologies with on-device AI capabilities, enabling conversational AI experiences across Indian languages without dependence on cloud infrastructure.

Launch of VYOMA Innovation Challenge

• Organizers: Digital India BHASHINI Division (MeitY), in collaboration with Current AI and Kalpa Impact.
• Objective: Promote open‑source, multilingual, voice‑first AI that functions in low‑connectivity or offline environments.
• Foundation: Builds on Sunno Sutra, a handheld multilingual AI device unveiled at the IndiaAI Impact Summit 2026.

Key Features

• Multilingual AI Infrastructure: Supports 36 Indian text languages, 23 voice languages, and 35 international languages.
• Offline Capability: Delivers conversational AI without cloud dependency, ensuring accessibility in rural and remote areas.
• Public Impact: Functions as digital public infrastructure, enabling inclusive access to services across India’s linguistic diversity.

Participation & Incentives

• Who Can Apply: Startups, MSMEs, researchers, students, academic institutions, and independent innovators.
• Support:
  • 20 shortlisted teams will receive developer kits and mentorship.
  • Finalists will present prototypes to an expert jury.
• Rewards: Winning teams eligible for ₹80 lakh in prizes and deployment opportunities with central and state governments.

Potential Applications

• Education: AI‑driven multilingual learning tools for schools in rural India.
• Healthcare: Voice‑enabled diagnostic and advisory systems in local languages.
• Agriculture: Offline AI assistants for farmers, offering crop and weather guidance.
• Governance: Citizen‑centric services accessible in native languages.

Strategic Impact

Focus Area	Benefit	Global Relevance
Digital Inclusion	Expands access to AI in low‑resource settings	Model for emerging economies
Multilingual AI	Supports 36 Indian + 35 global languages	Strengthens cross‑border collaboration
Offline AI	Works without internet/cloud	Critical for rural connectivity gaps
Open Source	Encourages innovation ecosystem	Aligns with global open‑tech movement

Challenges

• Hardware Optimization: Devices must be smaller, more efficient, and field‑ready.
• Scalability: Ensuring solutions can handle millions of users daily.
• Collaboration: Requires strong partnerships between startups, academia, and government.

Global Positioning

• India’s VYOMA Challenge positions the country as a leader in multilingual, offline AI innovation.
• Provides a blueprint for inclusive digital ecosystems worldwide.
• Reflects India’s ambition to make AI a public‑impact infrastructure, not just a commercial technology.

Weya AI Launches ‘Hush’: Lightweight Open-Source Speech Enhancement Model for BFSI Voice AI

weya AI, a BFSI-focused AI company building omni-channel AI agents for customer onboarding, sales, and collections today announced the release of Hush — an open-source speech enhancement model purpose-built for the realities of production voice AI. Trusted by Tier-1 institutions including Kotak Bank, Weya AI is on a mission to make AI transformation accessible across the global BFSI landscape through an on-premises voice AI stack.

At just 8 MB in size and requiring no GPU, Hush processes audio in under 1 ms per 10 ms frame and has been trained on over 10,000 hours of mixed data. The model is fully language-agnostic, with 1.8 million parameters, and is capable of operating consistently across all spoken languages. At launch, Hush ranked #5 on Hugging Face’s Audio-to-Audio leaderboard, making it one of the top-performing open-source models in its category.

Voice AI systems such as phone agents, call centre bots, real-time transcription pipelines, and conversational assistants often fail in real-world environments due to poor audio input rather than limitations of language models. When multiple speakers are present, traditional noise suppression systems either capture unwanted voices or degrade the clarity of the primary speaker, leading to unreliable outputs. This is one of the primary reasons voice AI fails in production.

Hush addresses this challenge by isolating the primary speaker from live audio streams while suppressing competing voices, background noise, secondary speakers, whistles, hum, hiss, and all other disturbances in real time. Built on the DeepFilterNet3 architecture and enhanced with an Auxiliary Separation Head, the model has been trained with competing human voices present in 60% of its dataset at signal-to-interference ratios of 12–24 dB.

Commenting on the development, Mr. Atul Singh, CTO, weya AI, said “Hush solves one of the most overlooked failure points in production voice AI. We built this because we kept seeing high-quality language models fail in the field, not because of the model, but because of the audio it was receiving. This is the first of several models we are developing internally, all oriented toward a single vision: giving enterprises —> banks, financial institutions, and regulated industries the ability to deploy world-class AI entirely on-premises, with full control over their data and infrastructure.”

Designed for seamless integration, Hush runs entirely on CPU and can be deployed across Linux, macOS (Apple Silicon), and Windows using prebuilt ONNX binaries, eliminating the need for heavy production dependencies. The model weights and full source code are available on Hugging Face and GitHub under the Apache 2.0 licence.

Voice Is the Next Frontier of AI

Voice technology has rapidly evolved from simple command-based assistants into sophisticated conversational agents. In 2025, the global voice AI market is valued at $6.8 billion and is projected to soar to $32.6 billion by 2033, growing at a CAGR of 18.4%. Already, more than 8.4 billion voice assistants are in circulation worldwide, signaling that voice has become a mainstream interface for human-machine interaction.

Why Voice Matters

• Natural Interaction: Speaking is faster and more intuitive than typing, making technology accessible to wider demographics.
• Context Awareness: Advances in natural language understanding allow voice AI to interpret intent, emotion, and nuance, moving beyond rigid command structures.
• Multimodal Integration: Voice is increasingly paired with vision (AR/VR, smart devices), creating seamless, immersive experiences.
• Business Transformation: Enterprises are deploying voicebots for customer service, healthcare diagnostics, and financial services, reducing costs while enhancing engagement.

Market Momentum

• Explosive Growth: Another report projects the voice AI ecosystem to expand from $3.14 billion in 2024 to $47.5 billion by 2034, at a staggering 34.8% CAGR.
• Sector Adoption: Banking, healthcare, retail, automotive, and entertainment are leading adopters, embedding voice into everyday workflows.
• Voice Biometrics: Security applications are gaining traction, with voiceprints used for authentication in financial and enterprise systems.

Challenges Ahead

• Accuracy Across Accents: Global adoption requires robust recognition across diverse languages and dialects.
• Privacy Concerns: Always-on microphones raise questions about surveillance and data misuse.
• Deepfake Risks: Voice cloning technology threatens trust, requiring stronger safeguards.
• Cultural Adaptation: Voice AI must respect local idioms and social norms to feel natural worldwide.

The Future of Voice AI

Voice is not just another interface — it is the bridge to ambient computing, where technology fades into the background and interaction feels as natural as conversation. From smart homes to healthcare diagnostics, from financial services to entertainment, voice AI is poised to redefine how humans engage with machines.

Voice AI Startup, SuperBryn Secures $1.2M in a Pre-seed Round Led by Kalaari Capital’s CXXO Initiative

Dr. Neethu Mariam Joy (Left) and Nikkitha Shanker (Right)

• With this funding and selection for the Mphasis Sparkle Innovation Program (US) via Nasscom InnoTrek 2025, SuperBryn will fast-track its Evals, Observability & Self-Learning layer to drive the next era of reliable enterprise voice AI.

SuperBryn, the Evals, Observability, and Self-Learning layer for enterprise voice AI, has raised $1.2 million in pre-seed funding led by Kalaari Capital CXXO Initiative, with participation from angel investors including Rikant Pitti (Co-founder, EaseMyTrip), Arjun Pillai (Founder, Docket AI), Sharath Keshava Narayanan (Founder, Sanas AI), Harish Manian (Group CEO, BMH), and Nivin Pauly (Leading South Indian Actor).

The funding will accelerate product development, expand engineering hiring, and deepen market validation with early enterprise customers across industries where dependable voice automation is now a strategic priority. Voice AI is a $47 billion market growing at 35% annually, but without a robust reliability infrastructure, much of this potential never reaches production, a gap SuperBryn is purpose-built to solve.

Commenting on the Fundraise, Nikkitha Shanker, Co-founder of SuperBryn, said: “Voice agents fail silently. An enterprise might have a million conversations a month, but they have no idea which ones went wrong, why the agent fumbled, or how to fix it without manually reviewing thousands of calls. We're building the layer that surfaces what's breaking, why it's breaking, and automatically makes the agent better, without human intervention. Monitoring and Evals is non-negotiable in industries like healthcare, finance, and insurance, where one failed conversation can mean a missed diagnosis, a compliance violation, or a claim that never gets processed."

Dr. Neethu Mariam Joy, Co-founder of SuperBryn, added: “After 14 years in speech and voice AI research, I’ve seen why voice agents fail in the wild. Most platforms test only for narrow conditions, not for the messy reality of human speech. SuperBryn exists to fix this. We’re building intelligent evaluation and feedback systems that ensure voice agents not only work on day one but keep improving every single day.”

Jayraj Bharat Patel, AVP, Kalaari Capital, said: “Voice AI is at an inflection point, enterprises are moving from experimentation to scaled deployment, but reliability remains the biggest bottleneck. SuperBryn will fill a critical missing layer with independent evaluation, monitoring, and continuous improvement. Nikkitha and Neethu have deep technical and 0-to-1 experience, and are extremely passionate about setting the reliability standard for voice AI globally. We are super excited to partner with them on this journey.

SuperBryn was founded nine months ago by Nikkitha Shanker, a second-time founder and NIT Calicut engineer, and Dr. Neethu Mariam Joy, a voice AI researcher with a PhD from IIT Madras and postdoctoral work at King’s College London. Founded by two women technologists from Kerala, SuperBryn emerged from a clear insight, voice agents excel in pilots but fail in production, especially with accents, noisy environments, multi-turn dialogue, and edge cases current platforms miss. Seeing the pattern, they set out to fix it by building the reliability layer that makes enterprise voice AI scalable.

Today, over 70% of pilots fail to reach production due to reliability issues in real-world environments. SuperBryn closes this gap enabling enterprises to move to production 20x faster and 10x cheaper, with early customers seeing resolution rates rise from under 40% to over 80% within 60 days.

These capabilities enable enterprises especially in healthcare, financial services, insurance, and other high-stakes environments to deploy voice AI with confidence and independent verification.From launch, the company is targeting global customers, with a strong early emphasis on the US market. SuperBryn’s selection as one of five startups for Nasscom's InnoTrek USA 2025 and its induction into the Mphasis Sparkle Innovation Program USA provide it with direct access to global enterprise clients, and the company is already working with US-based customers. SuperBryn’s long-term vision is to become the independent watchdog and reliability standard for enterprise AI agents, ensuring that voice AI systems meet the same monitoring, compliance, and reliability expectations as modern cloud infrastructure.

MyBox becomes only Indian Firm to Join Voice AI History With the Amazon Voice Interoperability Initiative

MyBox Technologies is proud to be the only Indian company to join Amazon's Voice Interoperability Initiative, a global consortium of like-minded companies building and promoting voice services on multiple products. This initiative was announced to the world by Amazon through their press release on 24th September 2019.

Through this initiative, MyBox Technologies will be able to work with leading companies that are part of this unique consortium and continue building on the rapidly expanding voice based ecosystem.

"A path breaking initiative which will change the way voice will spread across the world enabling people to talk with their devices and create ease of living. We are humbled to be considered as the only Indian company in this grand consortium. We hope we can contribute to this initiative and help make it a success in times to come," said Amit Kharabanda, Managing Director of MyBox Technologies.

MyBox is deeply committed to pushing boundaries in its sustained efforts in research and development to create unique voice based solutions. Solutions from MyBox allow operators to offer convenient and cost effective Alexa voice experiences to their customers. Multiple voice services will give the users a choice and interaction they have not had before and without being locked down to any particular system.

As a system integrator for Alexa voice services, MyBox will be able to extend these solutions to other OEMs across the globe. MyBox, a set-top box manufacturing company, was acquired by Hero Electronix, a venture by Hero group for new technologies space.

Launched in India in 2017, Alexa announced in September that it is able to understand and pronounce names of popular places, names and songs in regional languages like Hindi, Tamil, Telugu, Marathi and Punjabi but supported commands only in English.

~ Business Wire India