Gliglish Founder Interview
How Fabien launched a tool to learn languages with AI in 37 days and grew it to $8K MRR.
1. What's your startup's name, what does it do, how long have you been working on it and (optional) your MRR?
Gliglish is a new way to learn languages by speaking with AI.
You just speak into your mic and enjoy a conversation with a language teacher. Or you can roleplay a real-life situation, like buying a croissant at the bakery.
It's been exactly 16 months and revenue is now at €8K.
2. What's your name and what were you doing before?
I'm Fabien Snauwaert and I dropped out of college to play video games. Then I got a low-paying IT job and started working on an ebook on the side. Revenue was then between 30 and 500 bucks per month, for about 5 years.
Then I created a few books and courses and was able to go full time. But things got really interesting once I created a webapp for people to learn English. It made over €560K in total revenue… but that's over the course of 9 years.
Gliglish is the first product I build outside of the French market.
3. How did you come up with your startup and its name?
I love alliterations and rhymes. There are products (not mine) named Youglish and Gymglish. Then I made a product named Earglish, then another named Listglish. So then I was like, fuck it, let's just buy Gliglish.com.
It's hard on the tongue but easy to recall, I think? At least it's short and unique! That's in contrast to every other language app named SomethingLingua. But Pieter Levels insists I should change name, so there's that haha.
4. Are you a programmer, if so, how did you learn and how long did it take?
I was making HTML pages when I was 16, with a bit of vanilla JavaScript. And then I just learned everything on the go. Got a question? Google it. Or now ChatGPT/Cursor it.
I only got serious about it when I built the first webapp. Though I do have stuff online that's been running since 2007.
I learned design from asking friends and reading the excellent Refactoring UI by the Tailwind CSS guys.
5. How long did it take you to build an mvp, did you do any validation?
Took me two days to get a basic prototype. It didn't really work but was enough to get a sense that it could. Then it took me 37 days to make it work, make it fast enough, tweak, get WebSockets to work, deploy, add payment, and other boring details.
I launched like an idiot 2 days before Christmas, because I was broke af.
I have a newsletter for French people learning English. I always use it to launch products. It's my financial safety net.
6. What was your launch like, what did you do, was it a success?
People were enthusiastic about it, but the product didn't work so well just yet. I was clear I was still working on it though. It made only €615 in the first 30 days, so not exactly a homerun. But it felt better to get paid and get feedback than work in a bubble forever.
7. How long did it take to get your first customer?
It's been a rollercoaster:
- Got 4 sales on the day the product came out.
- Got zero sale when I launched on Product Hunt. I'd been agonizing the launch, and all this effort and worry was for nothing.
- People started talking about Gliglish once I posted on HackerNews. I was #2 on the site for a split second.
- Then I started getting traffic from AI directories and people into AI on TikTok & Instagram and that's when things started going viral.
- Then the site went viral in India with a video that got 2.4M views in 2 days and all hell broke loose. My Heroku server crashed and could not scale up, hitting all sorts of limits, including Postgres. I suspect I even caused a Heroku outage lol. But all this traffic was just vanity in the end, because people from India don't buy.
Eventually revenue kept growing and, right now, half of it is from B2C, with the other half coming from a partnernship I can't talk about. I have a few deals in progress.
Revenue goes up when I add new features people requested.
Right now people love the product but there's still a lot left to make the vision come true.
8. What's your top marketing methods and how can others start doing them?
That newsletter for French people learning English definitely allowed me to make rent a few times.
Then letting people try the product really easily with freemium is a mix of good product and marketing, I think.
Remember: as a user, you don't want sites to annoy you with a paywall. Being nice seems to just kinda work.
I don't market the product. It's all organic. So make a product worth talking about?
9. Whats the top 5 tools you use to run your startup?
- Markdown / Sublime Text for backlog, notes, TODO lists
- CursorAI to code
- Azure for hosting
- Front for email support
- SimpleAnalytics
oh and Spotify for morale
10. What would you do differently if starting over?
Chill about ProductHunt. It's a bit of a lottery and getting less and less relevant.
Talk to customers sooner and faster. Just grab someone on Facebook and go for coffee. Shut up and watch them use your product.
Prune bullshit and distractions like a madman, ESPECIALLY once you start getting traction. A lot of people are going to wanna talk to you with ZERO respect for your time. You need to be merciless: you'll be presented with good opportunities, but don't let yourself get caught in the bad ones (which is 80%+ of them.)
Iterate more: any success I had didn't come from waiting for years, it came from going for new products I was excited about that felt like they OBVIOUSLY were the future. It took me 17 years to ship 9 products, it's a mistake. If I'd iterated faster (instead of dragging too long on/between projects) I'd have gotten to the nice results faster.
Two final tips I would give my younger self:
- Only one thing matters: are you making your product, every day, something people want?
- Be brave! That's how you make good decisions and live by them. Fear robs you of opportunities. You will still feel the fear, but being brave shields you from it.