@dominic60184
Dominik LubańskiDeveloper▲11
Full Stack & AI Developer, AIvanta
2h ago · AI Engineering · Warsaw

Your AI API bill tells you what you spent. It doesn't tell you what actually made you money.

I've been building software for 13 years, and one thing I've learned is that costs become a real problem when you can't explain where they're coming from.

That's why I built AI API Cost Tracker, an open-source tool for understanding and controlling AI costs at the business level.

Langfuse, Helicone, and LangSmith are great for observability. But I wanted something more focused on the numbers that actually matter when you're running an AI-powered product.

  • Know exactly what each customer, feature, or agent run costs.
  • See which customers are profitable and which are quietly eating your margins.
  • Track real API costs, including cached tokens and pricing changes.
  • Catch budget overruns before they turn into expensive surprises.
  • Keep everything self-hosted, without exposing prompts or responses.

It's not another observability platform. It's a practical way to connect AI usage with real business costs.

Built for developers who ship AI products, not just experiment with them.

It's open source, and I'd genuinely appreciate feedback from anyone dealing with AI costs in production.

github.com/Dominic6018/ai-api-cost-tracker

Comments1

  • @ryanallis
    Ryan Allis
    Founder of AI Lounge and SaasRise, helping founders reach $100M+ exits
    48m ago

    The screen that shows which customers are quietly eating your margins is the one a SaaS founder would want to see before installing anything. Could you post a screenshot of that view here, even with made-up customers, so people can see what they'd actually be looking at each month?

Log in to join the thread.