General
Share and discuss tech, products, business, startups, or product recommendations
Start new thread

Would you vibe-code your own tools to avoid paying subscriptions?

by
•

A platform ( canivibecodeit) recently went viral on X for showing which paid tools you could realistically rebuild yourself.

Design tools, ChatGPT, or Claude are obviously hard to replace. But invoice tools, forms, simple CMSs, or dictation apps seem much more realistic.

The question is whether the build time and maintenance are worth saving $9–19 per month.

Which tools would you actually build instead of paying for them?

227 views

Add a comment

Replies

Best

It feels like choosing between renting versus owning an apartment. In a rented one you cannot change much, but fixing it is not your problem. In your own you get the flexibility and you also get every repair, plus nobody can raise the rent or sell the building out from under you.

We built our own content pipeline instead of paying for a writing tool, mainly because we wanted it to work our way rather than to save the subscription. On the other hand we pay for email delivery, since building that would buy us nothing.

 That’s a great analogy, Anastasiia! I also think it makes sense to build one or two mission-critical tools exactly the way you need them. But not at a scale where maintaining your own software stack becomes the main focus instead of doing the actual work.

What type of content is your writing tool built for? Is it focused on a specific social platform, blog posts, or something else?

And what other tools would you consider building yourselves rather than continuing to pay for subscriptions?

 It's for our blog posts.
I was thinking about vibe-coding a simple CRM, but a spreadsheet is more than enough for now. Salesforce and HubSpot offer free tiers, so money was never the reason. I just don't want the data sitting on someone else's server.

 That makes total sense. Privacy is a much better reason to build your own than saving a few bucks.

 What would you build for yourself instead of paying for a tool or service?

I don't think it's really about saving $9-19/month. It's more about investing some time upfront to build a tool that's perfectly integrated into your own workflow and can be reused whenever you need it.

I really like this fundamental approach: spend time once to create a convenient toolset, then work faster, avoid recurring costs, and have something that fits your process better than a generic off-the-shelf solution.

 I agree to an extent, Adana! But if you tried to do this for all 50 tools you use, it would make no sense. The time investment would be enormous.

For one or two very specific tools that are central to your workflow, absolutely. But rebuilding your entire software stack just to avoid subscriptions would probably cost far more in time than the subscriptions themselves.

Which tools would you personally rebuild instead of continuing to pay for them?

 Not all of the stack, of course =)

For example, we built an agentic CMS that handles a big part of our workflow. It analyzes competitors, automatically publishes blog content to custom domains, generates and maintains live documentation directly from developers' prompts, and helps us create complete web pages and subpages.

Those are the kinds of highly specialized workflows that are worth building in-house because they give us exactly what we need.

 This is really useful! Do you plan to stay inhouse or planning to share with others as well?

 We're planning to launch it on PH.

The interesting part is that while building , we ended up building this CMS along the way, and realized we'd accidentally created a second product ))

Now we're polishing it up, because we think it has value well beyond our own internal use.

100%, I vibe coded my own version of Wispr Flow, it barely costs anything in tokens, works just as well, and I tweaked some little things that annoyed about the UX of the original. Not sure I'd vibe code anything handling financial data, though, like invoicing software

 Cool, Denitsa! Really cool! How much does it cost you per month compared to Wispr Flow? I use some open-source tool for that, although building my own might actually make sense too.

Is there anything else you think you’d vibe-code for yourself?

 about $0.50 per month, I use ElevenLabs for voice-to-text and a Gemini model for cleanup (and "text tightening" if needed). Wispr Flow is $15. Sadly, all the other subscription tools I use would be virtually impossible to vibe code (stuff like Canva and CapCut for example)

 wow, this is worth it! Might be great to vibe code for me. Can you share the exact process to build it?

Yeah, some tools like Canva/Figma/ElevenLabs...not possible

 honestly, wish I came up with it, but I just followed this one tutorial

+ then asked Claude Code to add functionalities I specifically wanted (like a voice command to tighten and improve the text) that weren't in the tutorial

Sometimes I do it yea, not all the time tho

 nice

can you name some of the things you vibe coded for yourself?

 Most of them bite the dust unfortunately. Took a pause in the Indie world now I've comeback, so yea can't really give any examples as I forgot them 😆

we actually did this for a few internal tools. the honest math is that $15/month saves you maybe 2-3 hours of build time, but maintenance is the hidden cost nobody mentions. every time something breaks at 2am you're paying for that "free" tool in a different currency. i'd vibe-code anything that touches your own data pipeline or has a simple, stable use case. anything that depends on third-party APIs or needs constant updates, just pay for it.

 nice to hear your perspective!! I guess this is the main issue. It's easy to build, it's hard to maintatin (you should at least check it out).

Can you name of your best vibe coded tools that replace the subscriptions products

Love this thread! I think about that a lot: what will be the economic value of apps in the era of AI and vibe coding? I believe there are still a lot of barriers to vibe-coding and sustaining certain types of apps, especially ones with more robust backend architecture (embedding higher domain knowledge/intelligence), great UX, and integrations. Apps that are not integrated, without any higher level of domain knowledge around the problem they try to solve, and performing more mechanical tasks can more easily be vibe-coded.

 agree, it's way easier than before to build an app but it's still not super easy. How do you think will be solved the integration part?

, I believe the value for apps will continue to be in the domain knowledge the app holds (what logic does it use to solve a given problem, and how hard would it be for someone to replicate it? AI may help with this, but it still needs great instructions and context — without that, it will just produce mediocre outputs and underdeliver). Integration is also a tough one, in my opinion, and incumbent players will leverage this as a moat to stay relevant. I believe this will get harder in the future, as the criteria to integrate new apps/partners will increase (I am already facing this challenge myself… haha). It will represent a barrier for someone to vibe-code replicas of existing apps. You may do it, but without integrations and automated data flows, the value may not be there, and you may still be better off paying for existing solutions.

I do this systematically. I call it marketing engineering. Instead of renting a SaaS stack, I build the pieces I can and integrate only what I can't replicate. Fluxr itself started as a personal tool I built because nothing on the market did exactly what I needed. The line I draw: build if it gives you a competitive edge, rent if it's commodity infrastructure.

I've vibe coded a lot of tools for my start-up. Some of the tools I need don't seem to exist. I also prefer vibe coding others, because it make it easier to integrate data. I recently launched my third start-up and vibe coding has transformed the way I work.