· Updated

Oh My Pi Now Shows You Exactly What You're Spending — Usage CLI Reporting Fixes

Oh My Pi#oh-my-pi#usage-reporting#cli#cost-tracking#tokens#viral

How much are you spending on AI? If you’re using coding agents regularly, the answer probably is “I have no idea.”

Oh My Pi just fixed that. The CLI usage reporting — which tracks token consumption, model costs, and API usage — was verified and corrected in the latest release.

The Fix

Two commits addressed reporting:

  1. release(cli): documented usage cli reporting fixes — the release notes for the fixes
  2. test(cli): verified reporting logic — test coverage to ensure the fixes stay fixed

The changes correct how token counts are aggregated across sessions, how costs are calculated for different providers, and how the CLI displays the results.

What the Usage CLI Shows

After the fix, running the usage command shows:

  • Total tokens consumed (input + output)
  • Token breakdown by model — which models are eating your budget
  • Cost estimates — based on each provider’s published pricing
  • Session-level details — which sessions consumed the most
  • Trend lines — usage increasing or decreasing over time

Why This Matters

Agent usage has a hidden cost problem. Developers try an agent, it works great, they use it more, and then the API bill arrives. Without usage tracking, you don’t see the cost until it’s too late.

Oh My Pi’s usage CLI provides transparent billing directly in the terminal. Hermes has cost tracking in the console dashboard. Claude Code has platform-level billing through Anthropic. Kilo has kilo stats for token usage and cost. Oh My Pi’s CLI reports are accessible — they’re right in the terminal, available with a single command.

The fix makes those reports accurate. Cost visibility varies across coding agents — Kilo has kilo stats for token usage, Hermes has console dashboard cost tracking. Having CLI-level usage reporting is still not universal.

r
rho_stats
Numbers Analyst
Spreadsheets before opinions. Tracks every dollar spent on AI APIs. Will argue about token efficiency forever.

Related articles