Indie App Devs #7
Weekly tips for indie app developers.
Hello! 👋
Today we have another well-known indie developer - Anum Mian.
She shares how tackling a simple problem she faced led to a public launch after she shared it publicly.
Follow her on X/Twitter & LinkedIn.
When Feedback Shapes the Roadmap Better Than You Do
Unless you’ve been living under a rock, you might have heard of Sora, an app by OpenAI that allows you to create AI videos.
While AI video creation tools have been around for a long time, Sora was a game changer. Why? Because it was free to use for the general public.
This was incredibly exciting as an indie developer, since one thing we really struggle with is marketing. Sora opened up so many avenues for us. Previously, it took hours to create engaging videos for TikTok, Instagram, and YouTube. With Sora 2, it became possible to generate amazing videos instantly.
There was just one tiny issue: the Sora watermark.
Tackling the Watermark Problem
Like other developers, I created a bunch of videos and then Googled ways to remove the watermarks. However, using online tools felt like a lot of work, not to mention extremely expensive.
So the cheapskate that I am, I decided to quickly use FFmpeg to build a tool for removing watermarks locally on my machine. I whipped up a macOS application and added the capability to target the three common watermark locations in most Sora videos.
The result was decent: while some videos showed a bit of blur where the watermark had been, half were actually usable.
Viral Buzz on X
I have a habit of sharing what I’m working on X (previously Twitter), so I posted a screenshot of the tool on my timeline.
Everyone wanted it. I was overwhelmed by the requests, with the post reaching 160K views, over 400 bookmarks, and more than 500 likes. It was obvious people were begging for this tool.
Evolving from Personal Hack to Public Beta
The trouble was, I hadn’t created it for public use (it was a very unpolished product I’d thrown together for myself). But since demand was so high, I decided to release a beta version.
What bothered me most was the blur in some videos; FFmpeg did a decent job, but not a great one. I stayed up all night experimenting for a better solution. The obvious path was AI or a watermark remover API, but I needed something cheap and capable of bulk processing.
In the end, I found a way to use AI locally for watermark removal. Initially, it took 20 minutes per video (unusable, of course), so I optimized it for speed.
I launched the beta and posted about it on X. Within five minutes, a couple of users bought it. It was insane, though they quickly reported installation issues. I worked relentlessly with those first users to ensure smooth setup and a bug-free experience.
Iterating Based on Real User Feedback
I went through many iterations to make the app perfect for everyone. To keep things manageable, I limited support to M-series Macs, as Intel machines were too slow and problematic.
The best part?
All the helping users were paid ones, and the incoming revenue motivated me to polish it into a great product.
A few users suggested features I hadn’t considered. While most Sora watermarks appear in fixed spots, some wanted flexibility for other watermarks or varying positions. One developer user proposed letting people select custom locations, which I implemented along with other requests (all within a week, via several beta releases).
At that stage, I skipped analytics and auto-updates to focus on core features and fixes first, aiming for a reliable product later.
That approach paid off: I got amazing feedback from nearly every user.
Key Lessons Learned
What I learned from this was that building with real users helps you prioritize high-value features. I initially thought video upscaling would be a hit, but no one asked for it.
Instead, they wanted practical tools like metadata removal, custom watermark selection, and bulk uploading. By delivering exactly that, I created a product every user loved.
What’s Next?
Now, I need to generate more Sora videos to market this tool. Wish me luck!
Want to learn more?
Follow Anum on X/Twitter & LinkedIn as well.
I do admire the process, the technical solutions incorporated and you writing about this. Considering how good AI has become at creating videos and social media being flooded with AI videos, at the end I find it to be an unethical product. There is a reason for watermarks. Too many people are already fooled today by fake AI videos and this product makes this easier.