Startups developing voice AI technology and applications are having their moment. Model builders like ElevenLabs and Cartesia have raised millions of dollars in the last few months. Applications such as AI-powered notetaker Granola and meeting tools Read AI and Fireflies AI have also received investor attention and backing.
Continuing the trend, dictation app Wispr Flow announced today that it is raising $30 million in Series A funding from Menlo Ventures with participation from NEA, 8VC, Opal CEO Kenneth Schlenker, Pinterest founder Evan Sharp, Carta CEO Henry Ward, and Lindy CEO Flo Crivelli. Menlo Ventures’ Matt Kraning, who also backed the company as an angel investor, will join its board. To date, the company has raised $56 million.
The startup’s founder and CEO, Tanay Kothari, started building Wispr to create a device that would allow users to type just by mouthing words silently. Its prior funding was for that business.
Last year, the company instead started focusing on Wispr Flow, the software interface designed for the hardware device.
The company released a Mac app in October 2024, followed by a Windows app in March 2025 and an iOS app earlier this month. Kothari mentioned that, since its early release, VCs in Silicon Valley have been using the product.
“I think every single tier one venture fund in the valley uses Wispr Flow for their emails, memos, documents, and more. They feel themselves being hooked on it, and it is one of the products they use every day. Because of this, we started getting a lot of inbound,” Kothari said about investor interest.
Notably, Granola also had a similar story of receiving immense investor interest because VCs used their product a lot.
Startup’s CEO Tanay Kothari.Image Credits:Wispr Flow
Kothari also noted that the startup will soon achieve profitability at the current rate of growth, and initially, he didn’t want to raise money. However, he worried that Big Tech players with a massive distribution advantage could be a risk to the company. He wanted to rapidly multiply the company’s revenue and reach and decided to take the investment.
Menlo Ventures CEO Matt Kraning, who has been an avid user of the app, said that his initial thesis for Wispr Flow was that with the current set of input methods, like keyboards, we are “waiting for our thumbs to catch up with our thoughts.”
“Wispr Flow is creating an efficient way to translate digital thoughts and intent. The app captures users’ speech and what they want to convey very well. The team has thought about how people speak while developing models rather than focusing on things like word error rates,” he told TechCrunch.
User growth and future roadmap
The startup said that the app has been growing its user base by 50% month-over-month. Kothari noted that 40% of users of the app are in the U.S., 30% in Europe, and 30% in other parts of the world. In addition, more than 30% of the app’s users are from a nontechnical background.
“More and more people are using AI tools, but still, there isn’t a good interface for people who are not techies. ChatGPT-style interface is the most common one, and that was released three and a half years ago. We are building for all kinds of users so they don’t have to write system prompts to interface with AI,” Kothari said.
At the moment, Wispr Flow supports dictation in 104 languages. Kothari said that 40% of dications are in English, and 60% of them are in the rest of the languages, with Spanish, French, German, Dutch, Hindi, and Mandarin being the top languages.
The company will use the funding to grow its team of 18 with roles in engineering and go-to-market. It will also release an Android app and cater to enterprise users by setting up company-wide phrase context and support teams.
The startup is working on building Flow into a product that is akin to an AI-powered assistant that knows more about your personal context and helps you do everyday tasks like send messages, take notes, and set reminders. Plus, the company said it’s working with some AI hardware partners, without naming them, to power the interaction layer.
