AI Usage Board - Claude Code & Codex usage summary

    1

    Summarize your local Claude Code and Codex usage from session logs: tokens, estimated cost, 5-hour and 7-day windows, per-model breakdown, daily trend and output speed. Read-only, no network calls, no credentials.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    About this skill

    Ask your agent "how much have I used Claude Code this week?" or "which model used the most tokens?" and it answers from the session logs already on your machine.

    What it reports

    • Requests, input/output/cache tokens and estimated USD cost per account
    • Last 5 hours and last 7 days, plus a daily trend (up to 90 days)
    • Per-model breakdown and recent output speed (tokens per second)
    • A complete flag with diagnostics, so missing or unreadable logs are never reported as "no usage"

    How it works A single Node.js script (built-in modules only) reads ~/.claude/projects and ~/.codex/sessions, prints one JSON summary and exits. It makes no network calls, starts no server, writes no files and reads no credentials. Prompt and response text never appears in the output.

    Cost estimates Token counts are exact. Cost is calculated only from a prices file you provide (USD per million tokens); the skill never invents rates.

    Open source under MIT. A browser dashboard version is available on GitHub.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.

    Cost is an estimate from the prices you provide, not a bill; it does not reflect subscription plans, quotas or surcharges. Only logs on the machine where the agent runs are read. Log files whose modification time is older than the reporting window are skipped. Error messages are in Chinese.

    Open docs

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 5 days ago

    • Free with an Agensi account

    Software
    Node.js 20+

    Permissions

    File Scopes

    ~/.claude/projects
    (read
    only)
    ~/.codex/sessions
    prices
    JSON
    file
    passed
    with
    --prices

    Read-only. Makes no network calls, starts no server, writes no files and reads no credentials. Prompt and response text never appears in the output.

    Listed5 days ago

    What's inside

    Frequently Asked Questions

    The skill never guesses prices. Pass a prices file with --prices (USD per million tokens, keyed by the exact model names in the summary). Token counts are exact either way.

    No. The script only reads local log files and prints a JSON summary. It opens no network connections.

    At least one account could not be read fully, for example a missing directory or read errors. The agent tells you which account, so missing logs are never shown as zero usage.

    New in LLM Ops

    Free