Alexander Van Le built a VC-backed company and failed. Then, after a stint in gainful employment, he went viral, quit his job, and bootstrapped AI Chat Flow and Starpop. Now, he's bringing in a total MRR of $20k.
Here's Alex on how he did it. đ
I have a degree in ComSci and Finance, and ever since high school, I knew I wanted to build my own business.
I cofounded a FinTech company that got VC backing and everything, but it didn't work out.
We had four cofounders, zero revenue, and poor go-to-market execution at the time of our pre-seed. We were just an overall bad company. I was surprised by how "hot" VCs considered us. We had many real offers.
I've become a bit biased against the VC mindset since then. I realized a disconnect exists between what investors seek and what founders should focus on to improve their businesses. We spent too much time discussing terms, strategy, storytelling, creating slides, and too little time marketing and understanding our customers.
But it gave me more blood on my teeth to prove myself. And learned from it.
I now know to take a marketing-first approach to product building. And I prefer to use capital for scaling rather than survival, if possible.
After that, I worked for a few years as a management consultant and then as a software engineer, building projects at night. Most of them flopped.
I became more active on X. I posted every day for months. No one really cared about my tweets. I hovered around 100 followers with maybe 10-20 impressions per post. But then, one tweet changed everything for me. I went viral on Twitter. 3M views on one post. If you get over a million views on anything, I believe you have a business. So I quit my job to pursue entrepreneurship full-time.
And that tweet became AI Flow Chat.
Now, I'm focused on two companies:
Starpop is a platform for creating AI videos and images. We focus on helping e-commerce brands and agencies make ads.
AI Flow Chat is a visual AI canvas for working with text and image models. You can feed various data, such as PDFs, websites, social media links, videos, and images, to teach the AI whatever topic you want it to help you with. AI Flow Chat customers are typically influencers or marketers who use AI to help with their scriptwriting and copywriting in a visually organized manner.
I'm currently at about $20k MRR across both Starpop and AI Flow Chat.
After my tweet went viral, I knew I needed to build what people saw in the tweet. So, I built AI Flow Chat. The hard part was figuring out how to market it and what problem it solved.
For Starpop, we drew inspiration from existing software. The problem and solution were already clear.
It's important to consider as few things as possible when developing the initial MVP. Keep it as simple as possible and get it out as fast as possible. It took me two weeks to build and launch AI Flow Chat, and one week to build and launch Starpop. And for the latter, we sold our first subscription within that first week.
Currently, we are using PostgreSQL on Supabase, NextJS on Vercel, and Render for a long-running server.
In addition, we are using all major AI providers such as OpenAI, Anthropic, Google, Fal, etc.
Keeping the infrastructure simple for as long as possible helps you add features and fix bugs fast. But simplicity is subjective; it depends on the technologies and providers you've used before.
I chose the tech stack I had the most experience with, which allowed me to anticipate future problems. In hindsight, choosing a React-based framework was a brilliant choice for AI and web development in general. The ecosystem is large and rapidly developing, so if I need help with something, I can easily find an open-source repo or a blog post describing the process. AI's knowledge of React frameworks also helps.
My biggest challenge was figuring out how to get noticed.
My initial virality helped immensely by giving me credibility, but it wears off. I then had to figure out how to consistently engineer virality across all my marketing channels, rather than relying on luck.
I wish I had spent even more time marketing up front â and learning to market. I still feel I lack many marketing skills.
What I've learned is to focus on the funnel in this order:
Impressions
Conversions
Retention
This order is important to avoid over-engineering a part that won't move the needle. If no one knows you or your product, a good high-converting offer won't matter. If you have no users, a good product that keeps users also won't matter.
For both Starpop and AI Flow Chat, our initial GTM strategy was cold DMing people directly.
My cofounder, David, has a saying: "Do the unscalable," which makes sense initially when you only have time. Trading your time for fast impressions allows you to sell quickly and rapidly iterate on your marketing story.
One of our most successful moves was "community infiltration". We joined Discord and Reddit subreddits and posted valuable content. This could be a guide or video â content that barely mentions our product but demonstrates genuine thought. People then reached out to us. Those who DM'ed us from these posts had a high conversion rate, probably around 50%. These customers also provided valuable feedback, as they are the type of people who readily reach out if something is wrong or if they believe the product can be improved.
Reddit is an interesting channel because everything is transparent. You can figure out what your competitors are doing and replicate them. A good Reddit post's effect is long-lasting, bringing in traffic for months. I plan to double down on Reddit for both my businesses since I've found a style that gets results there.
Beyond that, we are starting to see success with paid ads. We're also beginning to experiment with LinkedIn, YouTube, and short-form videos. I haven't found any huge successes yet, but I'm learning.
I primarily apply marketing fundamentals when creating posts. This approach allows for experimentation and continuous improvement.
For every post, I try to include and optimize for the following:
A scroll-stopping hook
An engaging statement that prompts comments or likes
Valuable content that encourages saves or likes for later
A call to action that encourages follows or business inquiries
Some specific strategies that helped me:
Trend jacking: For example, if a new technology is blowing up, create an article or tweet about it, as people are generally interested in learning more.
Brand jacking: For example, use well-known larger brands as scroll-stoppers.
Controversial or polarizing statements: This is challenging to execute effectively, but aim to write content that people will strongly agree or disagree with. This generates comments from both sides.
Founder-distribution fit is the most important concept for indie hackers without VC funding. Without deep pockets to pay for distribution, you must master it yourself. Some excel at X, others at Reddit, TikTok, YouTube, Reels, SEO, etc.
While it's tempting to build first and market later, you must prioritize learning how to get impressions. Learn copywriting. Create engaging visualizations. And perhaps, learn video editing.
Identify which marketing channel fits you best. Find a sustainable and fun channel, because you will do a lot of it.
Following people who are one or two steps ahead of you will accelerate your learning. Start by copying them.
People in your niche are excellent resources because you can more easily apply their methods to your own work.
That applies to competitors, too â if you're just starting out, having competitors in your space is a blessing. Their marketing is inherently visible, so you can identify everything they did, analyze their strategies, and apply them to your own businesses.
For now, I want to grow my businesses to $100K MRR, which will allow me to take much bigger bets, both business-wise and personally.
Beyond this financial goal, I love the game of entrepreneurship, so I hope to continue doing this on a larger scale.
To keep in touch and follow my story, connect with me on X or LinkedIn. And check out Starpop and AI Flow Chat!
Leave a Comment
The shift from the VC mindset to 'marketing-first' product building is such a vital lesson. Itâs easy to get buried in the tech stackâespecially with a ComSci backgroundâbut as Alex points out, the best product in the world doesn't matter if you have zero impressions. Love the 'community infiltration' strategy; itâs the ultimate way to build a feedback loop while moving the needle on MRR.
The shift from VC to bootstrapping is interesting because it changes what you optimize for. VC = growth metrics. Bootstrapping = cash flow. Most founders who make that switch say the hardest part isn't building differently, it's thinking differently. A portfolio approach spreads risk too, which is the opposite of what VCs want you to do.
The funnel priority order hit home: Impressions â Conversions â Retention. We just launched an AI tool and made the classic mistake of optimizing for conversions (Google Ads) before even having enough impressions. Spent money on clicks from people who had no idea who we were.
Now shifting to the "community infiltration" approach â showing up on Reddit, IH, Twitter daily. No pitch, just being helpful. Way slower but feels more sustainable.
"Do the unscalable" is great advice. Trading time for fast impressions when you have zero budget makes more sense than throwing money at ads too early.
Really inspiring to see how you pivoted after a VC-backed project didnât work out and built something that now earns meaningful revenue. In your experience, what was the biggest mindset shift that helped you stay focused and consistent while bootstrapping to $20k/mo?
The "founder-distribution fit" concept really hit home. As someone building FontPreview (a free font testing tool), I've realized I can't just build and hope people find it. I have to actually show up where designers hang out.
Your point about community infiltration is spot on. I've started engaging in typography subreddits â not promoting, just helping. And yeah, the conversion rate when people reach out to you naturally is totally different.
One question: When you say "cold DMing people directly" â how do you approach that without coming across as spammy? Would love to learn more about your actual message templates or approach there.
The VC mindset thing hit home. Used to think raising = validation. Now I'm 4 products in, all built on weekends, and the ones that got traction did so from distribution, not from the product itself. Curious how much of Alex's $20k comes from the content flywheel vs organic product discovery?
Focus on simplicity and user experience really makes a difference. Iâm building a music site using LyriaâŻ3 that lets anyone generate AI-powered music easily, without complicated tools. If youâre curious about creative AI in a platform people will actually use, check it out:Lyria3 .im â would love to hear what the community thinks about AI music tools!
This is one of the most honest takes on the VC-to-bootstrap transition I've read. The line about "narrative-market fit" vs actual product-market fit is sharp â I've seen so many founders confuse investor enthusiasm with customer demand.
What stands out to me:
1. **The 2-week MVP** â Intentionally leaving bugs in to ship fast is counterintuitive but smart. You're right that technical assumptions change anyway, so perfect code early is wasted effort.
2. **"Do the unscalable"** â Cold DMs and community infiltration get dismissed too often. 50% conversion from people who DM you after valuable content is insane compared to typical outbound.
3. **Founder-distribution fit** â This is the key insight. Without VC marketing budgets, you HAVE to become your own distribution. Finding a channel that's "sustainable and fun" is the difference between burning out and compounding.
The mental shift from "growth at all costs" to "sustainable first" is huge. Prioritizing fun and motivation first, company second â that's a framework more founders need to hear.
Great story, thanks for sharing the journey.
The pivot from VC-backed to bootstrap is something I see more and more. I personally did đ We documented this exact journey on our podcast (Public SaaS Builders), interviewed founders who raised, burned out, then switched to bootstrapping and found peace. The mental shift is huge: from "growth at all costs" to "sustainable first". Curious what was the hardest mindset change was for him? For most, it's accepting slower growth but better sleep.
the "sustainable first" shift is huge especially after VC experience. when you've been optimizing for growth metrics for investors, switching to asking "is this actually healthy?" feels almost counterintuitive at first. but I think the sleep thing is legit â you make way better decisions when you're not constantly anxious about runway.
Hey Francois, Alex here!
I can totally relate to the shift from "growth at all costs" to "sustainable first".
I think the biggest mindset shift was learning to live for myself instead of the company. When I had VC investors, I kept thinking I had to sacrifice myself to grow the company.
If I feel like Iâm âsacrificingâ something, it usually means thereâs something Iâm not enjoying on a personal level.
Now I prioritize having fun and staying motivated first, company second. If Iâm not having fun, the company will probably end up poorly anyway.
Awesome!
Great example of building multiple AI products with recurring revenue â targeting lightweight, focused solutions is clearly the way forward in the evolving AI space.
2 weeks for AI Flow Chat MVP is insane. What did you cut to ship that fast - did you hardcode stuff you knew you'd rebuild, or was the scope just genuinely minimal?
Hey Andrey, Alex here!
I started AI Flow Chat as a hackathon project with a friend of mine. During the hackathon weekend, we had a working prototype of the main UI and backend. It was good enough to get a feel for whether the UX made sense or not.
The scope was very minimal. If youâre curious, hereâs what the first version looked like: /watch?v=uILzsXHCPI8 (add it to a YouTube URL. I canât post links here).
And yes, I took a lot of shortcuts in the code. I almost intentionally left bugs in. I knew there would be edge cases where things would break, but with zero users it didnât matter. The foundation was also super shaky anyway, because the technical assumptions you make at the beginning of a project often donât stay the same.
Most of the code has since been rewritten several times now and tons of code has been thrown away at the same time too.
This hits hard. The âhotâ pre-seed phase can be deceptive â external validation feels like product-market fit when itâs often just narrative-market fit.
Iâve noticed the same disconnect: investors optimize for optionality and upside, founders need traction and signal. Marketing and customer conversations compound; storytelling without signal just delays the truth.
Using capital for scale instead of survival is a powerful shift.
"narrative-market fit" is a pretty good word for it. Totally agree with your points here! đ
The shift from VC-backed pressure to a lean, profitable portfolio is a fascinating mental transition. Diversifying into multiple AI products seems much more resilient than betting everything on one 'moonshot' today. Are you using a shared internal framework for these tools to ship faster, or is each one a completely custom stack?
Hey Artyom, Alex here!
I'm using the same core for all of my projects. I used AI Flow Chat to bootstrap the code for Stapop which is the reason why I could launch after only a week. Otherwise, it probably would've taken me much longer.
I'm not sharing the code in a centralized repository or anything like that. It's more the same code & file architecture I'm using with the same libraries. Whenever I update tech in one project, I tend to do it for both. And so the advantage is, whenever I find a bug in one project, I can fix it immediately in the other one too.
The benefit is that there's almost no context switch because it feels like I know where everything is across both code-bases.