Tanay Kothari and Sahaj Garg met on their first day at Stanford and started Wispr in 2021 to build something neither of them could stop thinking about: a wearable that read silently mouthed speech and turned it into text. They spent close to three years on it. The hardware worked. What did not work was everything around it. Kothari has written that the obsession behind it went back much further, to watching Iron Man in Delhi at the age of 10, teaching himself to code that night, and building more than 50 apps in secret through school on one hour of allowed screen time a day.

In July 2024 he made the call. He killed the hardware product after three years of development, and the company went from 40 employees to 5 overnight. In his own account he walked into their 22,000 square foot office every morning afterwards and found a ghost town, with no product and no clear path forward. His summary of that moment: any rational person would have shut down.

They did not. Wispr had been building a second thing on the side, a voice to text layer originally meant to feed the hardware, and that became the whole company. Kothari says they launched Wispr Flow six weeks after killing the hardware, and he is explicit about why the speed mattered. Most pivots die because founders get stuck between the old thing and the new thing, so they gave themselves no time to second guess.

The public launch came on 30 September 2024, when Kothari and Garg posted Flow to Hacker News themselves. It was a Mac app for Apple Silicon: hold the Fn key, speak into any application, release, and the text lands where your cursor is. There was a free tier capped at 2,000 words a week, a Pro plan starting at $12 a month on annual billing, and a discount code for the Hacker News crowd. That post is also where the pre launch work shows up. Kothari thanked thousands of beta users who had been giving feedback over the previous few months, so the paid launch was not a cold start.

We date the first paying customer to that launch, because it is the day the paid plan opened. Measured from the July 2024 decision to kill the hardware, that is 91 days from a company with no product to a company selling subscriptions. Both ends of the number come from Kothari's own posts: the July 2024 date from his LinkedIn write up of the pivot, the 30 September date from the Show HN thread he and Garg wrote.

The product decision underneath the launch is worth pulling out. Instead of optimising word error rate, the number speech teams normally report, Wispr optimised what it calls zero edit messages: the share of dictations you can send without touching a single word. Kothari's launch post put Apple and Google dictation under 5 percent on that measure and Flow between 50 and 70 percent depending on the use case. That is the difference between a tool you check and a tool you trust.

The second decision was the target. In his own write up of the growth playbook, Kothari puts it bluntly: everyone builds voice demos for booking flights, and Wispr went after the 100 emails you write every day. Less exciting, much bigger.

From there the numbers move fast, and almost all of them are relative rather than absolute. A Windows app followed in March 2025 and an iOS app in June 2025. When Wispr announced a $30 million Series A led by Menlo Ventures on 24 June 2025, Kothari wrote that users were producing 72 percent of their characters through Flow across nearly 70 different apps, that people were speaking more than 100 million words a week through it, and that eight out of ten users who activated Flow were still using it six months later.

In October 2025 he wrote that revenue was growing 50 percent month over month. On 20 November 2025 he announced $25 million in new funding, bringing the total raised to $81 million, and led with the line that the round followed a 10x increase in ARR over five months, alongside 40 percent month over month revenue growth and teams at 270 of the Fortune 500 using Flow daily. In April 2026 he described telling his team to stop optimising revenue altogether and replacing ARR as the top line metric with current daily active users, on the logic that the habit comes first and the payment follows. That same month he wrote that usage in India had tripled in three months with no campaigns or partnerships, making it the second biggest market on its own.

By August 2026 Wispr had raised a $280 million Series B at a $2 billion valuation, taking total funding to $361 million, with more than 60 billion words written through Flow and over 10,000 enterprises using it.

One honest caveat on the revenue. Kothari publishes growth multiples generously and absolute revenue never. There is no founder statement anywhere of a specific ARR figure, so every number above should be read as a ratio from a base he has not disclosed. What is verifiable from his own writing is the shape of the thing: a three year hardware bet written off in a single month, a replacement product in beta six weeks later, paid plans live inside a quarter, and a company that never went back.