How Growati Started: A Client, 700 Unlisted Videos, and a Shelved Idea

I met Lokesh Kabra in September 2025, and Growati was not the reason.

I was building RoastNest on my own at the time and needed automation work done. Lokesh ran a service business writing software for clients, so I hired him. We ended up talking most days, and at some point I went through everything I was working on — including the things sitting on a shelf that I wanted to build and had not started.

Growati was on that shelf. He said he liked the idea and that he would want to work on it at some point. Then he built my automation, we finished, and nothing happened for a few months.

The part I would not have predicted

In February he came back and said he actually wanted to do it.

What made that different from the September conversation was that he brought a problem with him. One of his clients — Yoga Marga, a yoga school running out of New York — had roughly seven hundred videos sitting unlisted on their YouTube channel. Recorded over years, uploaded, and then stuck. Every one of them needed a title, a description and a thumbnail before it could be published, and nobody was going to do that seven hundred times by hand.

That is a genuinely large job and it was also, precisely, the thing on the shelf.

We built it while doing it

The order matters and I think it is the reason the product is shaped the way it is.

We did not build a platform and then look for someone to use it. We took the Yoga Marga backlog and worked through it, building the automation as we went, because the work had to get done either way. Every constraint in the product now — the review step, the templates, the brand profile — exists because that engagement made it necessary rather than because it seemed like a good idea in a planning document.

By the time the last of those videos was published, the product was more or less ready. The full version of that engagement is its own post, because the interesting parts are operational rather than founding-story material.

Then beta, then launch

After Yoga Marga we went looking for channels that were not a yoga school, because a product shaped by exactly one customer is a consultancy with extra steps.

We onboarded a small group — a podcast channel, a couple of music and music-production channels — and gave them the beta. That group is where most of the awkward findings came from, including the pattern about prolific channels having the worst metadata.

We launched publicly in July 2026 and finished the day in the Product Hunt top ten, which brought more people through the door in a week than the previous five months combined. The first paying customers arrived shortly after.

Then we spent the weeks after it fixing what broke. A launch spike is a load test you did not schedule, and a meaningful amount of our automation fell over on our own servers under it. That period was less fun than the launch and did more for the product.

What I would push back on in our own story

Founding stories get told backwards, and this one has an obvious tidy version: two people meet, one has the idea, the other has the customer, everything clicks. That is roughly true and it flatters us more than it should.

The honest version is that the idea sat unbuilt for five months while I was busy with something else, and it only moved because someone else turned up with a deadline attached. I do not think I would have started it alone. Plenty of people would read that as a lesson about co-founders; I read it as a lesson about how weak an idea is until something forces it to exist.

The other thing worth saying: our numbers are small. Ten-odd beta channels, 500-plus synced videos, first revenue in the low hundreds. That is a real company and it is not a large one, and I would rather write that here than have you find it out.

Where we are now

Improving the thing rather than expanding it. The current work is performance-driven metadata revision, which is the first feature we have built that Yoga Marga did not need.

If you want the version of this that is about the work rather than about us, read what publishing 700 videos actually involved.