TokenTrace

I've been building AI agents/automations recently (part freelance work, part just exploring the space), and I kept running into the same annoying problem: I had no single place to see how much I was actually spending across OpenAI, Anthropic, and Gemini for different projects. I'd have to check each provider's dashboard separately, or just wait for the bill and hope it wasn't too bad.

So I built TokenTrace — a simple dashboard that pulls in your usage across providers, shows spend by project, and lets you set a monthly budget with alerts before you go over.

What it does (v1, kept deliberately simple):

• Connect your OpenAI/Anthropic/Gemini API keys (read-only usage access, encrypted at rest)

• See spend broken down by project and provider

• Set a monthly budget per project, get email alerts at 50/80/100%

What it's NOT trying to be: a deep tracing/debugging tool like Helicone or Langfuse. Those are great if you need to debug individual LLM calls — TokenTrace is just the simple "what am I spending, and warn me before I overspend" layer, for people who find the tracing-first tools like overkill for that.

It's live and free to try right now: tokentrace.in

I'm early — no paying users yet, just trying to figure out if this is actually useful to people beyond me. If you're building with multiple AI APIs, I'd genuinely appreciate you trying it and telling me honestly: is this something you'd actually use? What's missing? Is the budget-alert-only angle too narrow, or does it hit the right note versus something more complex?

Happy to answer anything about how it's built too — happy to share stack/architecture if useful to anyone else building something similar.