DEV Community

Paw from Oz
Paw from Oz

Posted on

I built a local-first CLI to track what I'm actually spending on AI APIs

#productivity #cli #devtools #ai

The worst part of using AI APIs isn't the cost. It's not knowing what you're spending until the bill arrives.

I wanted per-project cost tracking — which project used the most tokens? Which model was the expensive one? The cloud dashboards don't tell you that.

So I built AICostTracker — a local CLI that logs every API call and breaks down costs by project and model.

Auto-logging (the easy way)

Install as a library and wrap your AI client — every call is logged automatically, no manual work:

npm install @ozperium/aicost-tracker
Enter fullscreen mode Exit fullscreen mode 
import OpenAI from 'openai';
import { track } from '@ozperium/aicost-tracker';

const openai = track(new OpenAI(), { project: 'myapp' });

// Nothing else to do — all calls now logged
const response = await openai.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'hello' }]
});
Enter fullscreen mode Exit fullscreen mode

Works with Anthropic too:

import Anthropic from '@anthropic-ai/sdk';
import { track } from '@ozperium/aicost-tracker';

const anthropic = track(new Anthropic(), { project: 'docs-bot' });
// All messages.create calls now logged automatically
Enter fullscreen mode Exit fullscreen mode

The wrapper is transparent — it never surfaces logging errors to your code, and adds zero latency to the API call itself.

See your costs 

aicost summary
Enter fullscreen mode Exit fullscreen mode 
  AI Cost Tracker — Summary
  ══════════════════════════════════════════════════
  Total entries:    12
  Total cost:       $0.4821
  Input tokens:     48,000
  Output tokens:    31,200

  By Project:
  ──────────────────────────────────────────────────
  myapp                $    0.3102  8 calls
  docs-bot             $    0.1719  4 calls

  By Model:
  ──────────────────────────────────────────────────
  gpt-4o               $    0.3102  8 calls
  claude-3-5-sonnet    $    0.1719  4 calls
Enter fullscreen mode Exit fullscreen mode

Manual logging (CLI)

If you prefer not to install as a library, you can log calls manually:

npm install -g @ozperium/aicost-tracker
aicost log myapp gpt-4o 1500 800 "generated README"
aicost summary
Enter fullscreen mode Exit fullscreen mode

Why local-first?

All data lives in ~/.aicost/usage.jsonl — one JSON line per call, greppable, portable, yours. No cloud account, no telemetry, no dashboard to maintain.

Pricing is baked in for OpenAI (gpt-4o, o1, gpt-4-turbo, gpt-3.5), Anthropic (claude-3.5-sonnet, claude-3-opus, haiku), Google (gemini-1.5), and local/Ollama (cost: $0).

GitHub: https://github.com/Ozperium/aicost-tracker

Also in the stack: AgentSpec for testing AI agent behavior, and quota for monitoring rate limits before they stop you.

Top comments (0)