Voice AI startup Gnani.ai has introduced Artha, an end-to-end sovereign AI stack designed for Indian enterprises and public institutions. The platform was unveiled by Vice President C.P. Radhakrishnan in New Delhi on 28 August 2026. It combines the startup's Evon v3.3 open weight language model with its Plexus agentic AI platform to help organisations retain control over data and technology infrastructure.
A Sovereign Stack Built for Indian Needs
Artha addresses three challenges facing Indian organisations adopting artificial intelligence: data sovereignty, the cost of deploying AI at scale, and the ability to work effectively across Indian languages and real-world use cases. The stack includes Evon v3.3, Evon v2.0 for enterprise reasoning, Prisma v2.5 for speech recognition and Timbre v2.5 for speech synthesis, alongside the Plexus agentic AI platform. Open weights for Evon v3.3 are available upon request on Hugging Face under an Apache 2.0 licence.
Language Efficiency and Technical Performance
Evon v3.3 is a 30 billion parameter model trained natively across eleven Indian languages on roughly two trillion tokens, with a focus on reasoning and Indic-language performance. The startup said the model consumes around 40 percent fewer tokens for Indian-language workloads than comparable models, helping lower compute costs. Gnani.ai rebuilt the tokenizer for Indian scripts and plans to expand the model family with 70 billion and 100 billion parameter versions while increasing language coverage to 22.
Early Enterprise Interest and Use Cases
Gnani.ai has started showing the model to select customers, and five of twenty enterprises from a recent customer meeting in Pune have begun building on top of Evon. Potential early use cases include underwriting, advertising and payments reconciliation, although no deployments of Evon have gone live yet. The company expects regulated workflows such as underwriting and payments reconciliation to initially retain human oversight while routine document and data matching tasks become more autonomous.
Deployment and Governance
Gnani Artha is designed to run inside an organisation's own data centre or virtual private cloud, which the company says helps meet data protection, banking and insurance regulatory requirements. The model uses approximately 3.5 billion active parameters on any given token, allowing it to fit on a single inference node rather than a large cluster. Guardrails, observability and audit logging are built into the platform from the first deployment rather than added afterwards.
Government and IndiaAI Mission
Gnani.ai also outlined government use cases for Artha, including multilingual grievance resolution, beneficiary scheme enrolment and direct benefit transfer failure resolution. The launch aligns with the Indian government's push for indigenous foundational models, and Gnani.ai was selected under the IndiaAI Mission in 2025 to build a multilingual real-time voice AI foundational model. In December 2025, the startup launched Vachana STT, a speech-to-text model for Indian languages trained on more than one million hours of real-world voice data.
Company Background and Funding
Founded in 2017 by Ganesh Gopalan and Ananth Nagaraj, Bengaluru-based Gnani.ai offers voice AI products for enterprises and government bodies. Its portfolio includes virtual assistants, real-time agent guidance tools and voice biometrics for fraud detection across banking, insurance, healthcare, telecom and government services. Earlier this year, Gnani.ai raised $10 million in a Series B round led by Aavishkaar Capital with participation from Info Edge Ventures.
Gnani Artha represents a coordinated effort to combine sovereign large language models, speech capabilities and agentic workflows for Indian institutions. The launch signals growing competition among domestic startups to address data residency, cost and multilingual accuracy in regulated sectors. While enterprise adoption remains early, the company's IndiaAI Mission selection and existing voice AI customer base position it to pursue production deployments across banking, insurance and government services.