Recently, I have been working on a small AI SaaS product called AI Remover.
The idea is simple: many creators, marketers, and small teams often have videos or images that contain unwanted text, subtitles, captions, or watermark-like overlays. Instead of manually editing frame by frame, I wanted to build an online tool that makes this workflow easier.
The project is now available here:
What AI Remover does
AI Remover is designed to help users remove unwanted text from videos and images online.
Current use cases include:
- Remove text from video
- Remove subtitles from video
- Remove captions from video
- Remove watermark-like overlays from video
- Remove text from image
The goal is not just to create another upload button. I wanted the product to feel like a real workflow: upload a file, let AI process it, preview the result, and download a cleaner version.
Why I built it
While working on AI video tools, I noticed that many users do not search for general “AI video editing” tools.
They search for very specific problems, such as:
- “remove text from video”
- “remove subtitles from video”
- “remove captions from video”
- “remove watermark from video”
- “remove text from image”
This made me realize that small, problem-focused AI tools may have better search intent than broad AI platforms.
Instead of building a huge all-in-one editor first, I decided to focus on one clear pain point: removing unwanted text from visual content.
Product challenges
Building this kind of tool is harder than it looks.
The difficult parts are not only about the AI model. The real product challenges include:
- Different types of text
Text inside a video can appear in many forms:
- burned-in subtitles
- social media captions
- stickers
- product labels
- watermark-like overlays
- image text
- moving text across frames
A simple “remove this area” feature is not enough. The tool needs to understand where the unwanted text is and reconstruct the background naturally.
- Processing time
Video processing can be slow, especially when the file is long or high resolution.
For a web product, speed matters a lot. If users upload a short video and wait too long, they may leave before seeing the result.
So one of my ongoing goals is to improve processing speed and make the waiting experience more transparent.
- Result quality
Users do not only want the text to disappear.
They want the background to look natural after the text is removed.
This is especially difficult when the text is placed over faces, moving objects, complex backgrounds, or detailed scenes.
- UX and trust
For AI tools, users need to trust the result quickly.
That means the homepage, upload flow, examples, pricing, and before/after previews all matter. A technically working product is not enough if users do not understand what it can do within a few seconds.
Tech stack
The project is built as a modern web application.
The current stack includes:
- Next.js
- Tailwind CSS
- Cloudflare
- Vercel
- AI video/image processing APIs
- Analytics tools for tracking user behavior and conversion
I am also paying close attention to SEO, because this type of tool depends heavily on search intent.
What I learned
A few things I learned while building this product:
Narrow tools can be powerful
A focused tool that solves one painful problem can sometimes be easier to explain than a large platform with many features.
“Remove text from video” is much clearer than “AI video editing platform”.
SEO affects product design
When building an SEO-driven SaaS, the product structure is not only about UI.
Routes, landing pages, examples, headings, and feature names all matter.
For example, pages like “remove subtitles from video” and “remove captions from video” may look similar, but users searching for them may have different intent.
AI quality is only one part of the product
Even if the AI model works, users still care about:
- upload experience
- processing speed
- preview clarity
- pricing
- free trial
- trust signals
- before/after examples
A good AI SaaS product is a combination of model quality, UX, positioning, and distribution.
What’s next
I am continuing to improve AI Remover in several areas:
- better video text removal quality
- faster processing
- better before/after examples
- clearer pricing
- more specific landing pages for different use cases
- better support for creators and marketers
This is still an early-stage product, but it has already taught me a lot about building practical AI tools for real user problems.
If you are interested, you can try it here:
I would also love to hear feedback from other developers and indie hackers building AI tools.
Top comments (0)