AI-Based Voice Platform MiHup Secures US$ 1.7 Million In Series A Round


MiHup’s offering allows businesses to add voice control to products as well as to automate quality assurance tasks

MiHup, an artificial intelligence (AI)-based voice assistant start-up, announced that it has raised US$ 1.7 million (Rs. 12.5 crore) in a Series A funding round from venture capital firm Accel Partners and Ideaspring Capital.

Commenting on the development, Biplab Chakraborty, co-founder and COO, MiHup, said, “This will help us to further our ambition of adding more Indian languages into our engine. We support English, Hindi and Bengali fully and very limited use cases for additional 12 Indian languages. Our immediate aim is to fully support the next eight most spoken Indian languages.”

The start-up targets automotive original equipment manufacturers (OEMs) and business process management (BPM) firms. The company’s offering allows businesses to add voice control to products as well as to automate quality assurance tasks, which are done largely manually today.

Suitable for car control use cases

The AI-enabled regional voice recognition platform supports a mix of languages and dialects. Currently, the service is available in English, Hindi and Bengali. The platform is ideal for use cases like car control, entertainment and media and customer call analysis.

Briefing about the use cases, Tapan Barman, co-founder and director, MiHup, said, “One of the use cases we’ve started working on is a virtual voice assistant for cars, which will be able to work offline as well. We’ve partnered with Harman Kardon for this and we’re working with two of the largest auto OEMs in the country currently.”

In 2016, MiHup had raised Rs. 45 crore in a seed funding round from Accel Partners, which took a 20 per cent stake in the start-up as part of the deal.

The start-up was founded by Biplab Chakraborty, Sandipan Mandal, Sandipan Chattopadhyay and Tapan Barman in 2016. Driven by intelligent voice interfaces, the company claims that its platform provides human-like understanding of naturally spoken questions for big, complex content fields.