DEV Community

Thales Horta
Thales Horta

Posted on

I built a tool to compare cloud and AI costs, because I couldn’t trust the numbers anymore

If you’ve ever worked with cloud infrastructure, you probably had this feeling:

“Are we overpaying for this?”

Now add AI into the mix — OpenAI, Anthropic, Gemini — and it gets even harder.


The problem

Most tools today focus on what you already spent.

But that’s not the real question.

The real questions are:

  • What would this workload cost on another cloud provider?
  • Is AWS really more expensive than OCI for this setup?
  • How much more am I paying by choosing one LLM over another?
  • What happens to my cost when I scale this?

Surprisingly, it’s still hard to answer these quickly.


What I built

I created a simple tool called:

👉 https://costifyhq.com

The idea is straightforward:

  • Input a workload (CPU, RAM, storage, traffic)
  • Or simulate AI usage (tokens, requests)
  • Instantly compare costs across:

    • AWS, Azure, GCP, OCI
    • OpenAI, Anthropic, Gemini

And get a quick answer like:

“You could save ~25% by switching providers”


Why this approach

I didn’t want another dashboard.

I wanted something that answers:

“What should I be paying instead?”

So the focus is:

  • Fast input
  • Simple estimates
  • Clear comparison
  • Immediate insights

Not perfect accuracy — but fast decision-making


Example

Let’s say you run:

  • 8 vCPU
  • 16 GB RAM
  • some storage and traffic

Instead of digging through pricing pages, you just get:

AWS:   $1,240
OCI:   $920
GCP:   $1,050

→ Potential savings: ~25%
Enter fullscreen mode Exit fullscreen mode

Same idea for AI usage:

OpenAI:   $120
Anthropic: $95
Gemini:    $80
Enter fullscreen mode Exit fullscreen mode

What I’m trying to validate

This is still early, and I’m trying to understand:

  • Is this something people would actually use in real workflows?
  • Would you trust a tool like this for decision-making?
  • What’s missing to make this truly useful?

Where this could go

Some ideas I’m exploring:

  • Saving and tracking cost scenarios over time
  • Alerts when a cheaper option becomes available
  • Real integrations with cloud billing
  • Smarter recommendations (not just comparison)

Would love your feedback

If you work with cloud or AI costs, I’d really appreciate your thoughts.

Even if it’s:

  • “This is useless”
  • “This is interesting but missing X”
  • “I would actually use this if…”

👉 https://costifyhq.com

Trying to figure out if this solves a real problem — or just one I’ve been living with.

Thanks!

Top comments (0)