vibestats/Guides/How to track Claude Code usage

Guide

How to track Claude Code usage

A practical guide to tracking Claude Code usage locally with vibestats.

Intent
Guide
Focus
How to track Claude Code usage is a practical workflow page for turning local AI coding usage into something readable and repeatable.

Highlights

  • Start with daily usage, then refine
  • Model and session views answer different follow-up questions
  • Diagnostics help when the totals look confusing

Relevant commands

npx vibestatsnpx vibestats --monthlynpx vibestats --sessions

Step 1

Run the default usage view

Start with `npx vibestats` to see the default daily report for Claude-compatible usage.

Step 2

Change the report shape

Switch to monthly, total, model, or session views depending on whether you care about cadence, totals, model mix, or outlier sessions.

Step 3

Add deeper context

Use activity, wrapped, or diagnostics routes when the table alone does not answer the question.

Why this guide exists

Most AI coding reporting problems are not about one missing command. They are about choosing the right surface: daily usage, wrapped summaries, activity heatmaps, cost views, or shareable aggregate pages.

What to expect

This guide stays focused on vibestats workflows and the pages already documented in the public command reference, so you can move from local data to a readable result quickly.

FAQ

What is the fastest way to approach how to track claude code usage?

Start with the relevant command, verify the output locally, then decide whether you need a share page, a wrapped summary, or a heatmap for communication.

Do I need to upload raw conversations for these guides?

No. vibestats works from local usage artifacts and only turns aggregate results into hosted pages when you explicitly publish them.

Related pages

Continue by intent

View all guides