Skip to content
← Tools / Skills
Toolccusage/10 trustVerified

ccusage

See exactly what your AI coding agents cost: daily, weekly, and monthly token usage reports for 18+ CLIs.

Visit official site →

About

ccusage reads the usage data your coding-agent CLIs already write locally and turns it into clean daily, weekly, monthly, and per-session reports. One command shows tokens in, tokens out, cache reads and writes, and estimated cost, broken down by model. It started with Claude Code and now covers Codex, OpenCode, Amp, Droid, Codebuff, Goose, Kimi, Qwen, GitHub Copilot CLI, Gemini CLI, and more, with a single aggregated report across every detected source.

Everything is local: it parses the JSONL and config files on your machine, so there is no account, no telemetry, and nothing to upload. At 19,000+ stars and heavy npm traffic it has become the standard answer to 'why did my API bill spike'. The limitation is inherent to the approach: it can only report what the CLIs record locally, so coverage depends on which agents you have installed, and cost estimates are only as good as the pricing tables it ships with.

What it can do

  • Daily, weekly, monthly, and per-session token and cost reports
  • 18+ supported sources: Claude Code, Codex, OpenCode, Amp, Droid, Goose, Kimi, Qwen, Copilot CLI, Gemini CLI, and more
  • Per-model breakdowns with cache read/write visibility
  • Aggregated report across every detected agent in one view
  • Fully local: parses on-disk usage files, no account or telemetry
  • JSON output for scripting and dashboards; live dashboard mode

Setup

  1. Run it with no install: npx ccusage@latest (or bunx ccusage, pnpm dlx ccusage, nix run github:ccusage/ccusage).
  2. npx ccusage@latest daily for today's report across all detected agents, or scope it: ccusage claude daily, ccusage codex daily.
  3. Use weekly, monthly, or session for other views; add --json for machine-readable output.
  4. Optional: install globally with npm i -g ccusage if you run it often.

Prerequisites

Node.js (for npx/bunx); at least one supported coding-agent CLI installed so there is local usage data to read.

Use cases

developer-toolsanalyticscost-trackingagentsopen-source

Works with

Trust score — 8/10

/10 trustVerified

Scored 8/10 (verified tier, metrics-v1). Evidence: 19,018 GitHub stars (GitHub API), MIT, pushed Oct 15 2026; the README documents the multi-agent source table, the daily/weekly/monthly/session report model, and the local-only parsing design; npm and ccusage.com corroborate the install story. Strengths: the standard usage-reporting tool for coding agents, zero-config, privacy-friendly. Docked for: (1) can only report what CLIs record locally; (2) cost estimates depend on bundled pricing tables; (3) young project, still evolving source coverage. License: MIT.

★ 19,018 GitHub stars

How the trust score is built

Full methodology →
Adoption · stars/downloads, log-scaled vs peers
8.0/10
Maintenance · recency, releases, issue load
10.0/10
Openness · license + docs
10.0/10
Manual review · sanity read + red-flag check
8.0/10
  • · Adoption 8: starScore from 19,018 stars on ccusage/ccusage plus heavy npm traffic (log scale vs 298k peer max)
  • · Maintenance 10: pushed 2026-10-15 (recency 10); moved to dedicated org, active development
  • · Openness 10: MIT (0.7*10) + docs site verified at ccusage.com (0.3*10)
  • · Manual 8: standard usage-reporting tool corroborated by npm and multi-agent source table; docked: local-only data limits coverage, cost estimates depend on bundled pricing

Sources

Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.