If LLMs can make that simple tool, make, don't buy
Like many indie hackers/small business owners who are trying to stay lean, I don't want to pay for many subscriptions (especially if I'm only using their basic features. I also get annoyed by very limited free plans (e.g. Loom).
So, I made a bunch of tools with the basic features I need + custom features I want:
- Screen recorder with two layouts and annotation for simple internal or demo videos (like Loom)
- Kanban board for content ideas (like Trello but with a text mode)
- Minimalistic writing app
- Visual bookmark manager for organizing a moodboard (like Raindrop)
- Read later app
- Bulk file renamer based on rules
- Subscriptions tracker (imagine paying a subscription for a sub tracker!)
It's funny because the top post today is Aytekin Tank's post, Should you build that internal ai tool — or just pay for it?. That is about AI tools specifically but in general because the LLMs are so good at coding now, many simple internal tools can be and should be hacked together quickly. All that $20/mo subscriptions add up quickly.
This isn't to say we shouldn't pay for any tools. I still pay for apps like ScreenFlow (for making complex launch videos), Claude, Figma, etc. I won't even dream of making them with AI.
But if something can be quickly vibe coded and solves a real pain that others might not understand, even temporarily, I think we can/should build, instead of buy.
Thoughts?
The subscription fatigue is real. That bulk file renamer is the perfect example - such a narrow use case that no SaaS would get right anyway. The trick is knowing where the complexity threshold sits. Once you need authentication or payments those quick hacks become expensive technical debt.