Bangalore-Based Murf AI Launches Falcon 2 to Compete in Voice AI Market

Bangalore-headquartered Murf AI has publicly launched its text-to-speech model, Falcon 2, aiming to compete with major players like OpenAI and ElevenLabs. The startup positions the new model as a lower-cost, high-speed alternative for enterprise applications such as call centers and customer communications.

Bangalore-based startup Murf AI has entered the competitive artificial intelligence voice market with the public release of its text-to-speech model, Falcon 2. Available starting August 20, the platform has recorded high scores compared to offerings from better-funded international competitors like OpenAI on independent tracking platforms such as Artificial Analysis.

Founded in 2020 by Indian Institute of Technology Kharagpur alumni Sneha Roy, Ankur Edkie, and Divyanshu Pandey, Murf AI is backed by venture capital firms Matrix Partners and Elevation Capital. The company currently serves over 10 million users across 300 corporate clients, including major global enterprises such as Pfizer, Cisco Systems, Honeywell International, Nestle, and Air France-KLM.

Developed entirely by the startup's India-based engineering team, Falcon 2 is designed to balance natural-sounding speech generation with low latency and economical pricing. The model responds in under 100 milliseconds and is priced at $0.01 per generated minute, which amounts to $10,000 for every million minutes. This positions the offering at a significantly lower price point compared to competitors like ElevenLabs Turbo and Flash, which are priced at $50,000 for the same volume. The model supports over 150 voices across more than 35 languages.

Speaking on the broader industry landscape, Sneha Roy, Murf AI’s chief operating officer, noted that major AI models have historically emerged primarily from the United States. Falcon 2, she stated, is engineered to compete directly with established global models while meeting the specific economic and deployment requirements of enterprise production.

The training data for Falcon 2 incorporated real-world interactions, including live customer calls and booking flows, rather than relying exclusively on scripted narration. The system is built to interpret sentence context and dynamically adjust pace, emphasis, and pronunciation. Murf AI stated that the platform improves voice delivery on details that frequently challenge AI systems, such as currencies, dates, addresses, and license plate numbers. The infrastructure can handle up to 10,000 concurrent calls.

To address data security and compliance needs for large enterprises, Murf AI provides cloud support across 11 geographies and offers deployment options within in-house enterprise clouds. Regarding voice talent acquisition, Roy stated that the company works with consenting professional voice actors through transparent contracts, ongoing royalties, and data protection measures.

"The expansion of Indian startups into core deep-tech and artificial intelligence categories demonstrates the maturity of our engineering talent. Murf AI's entry into the global voice generation market highlights a distinct shift toward cost efficiency and scale in enterprise tech. By optimizing pricing and latency for high-volume users like call centers and banks, the company addresses a clear commercial demand. Competing directly with well-funded global players will require continued technical resilience and sustained enterprise adoption." — Dr. Shishir Gupta, Founder & CEO, StartupLanes