AI Token Monitor

What AI Token Monitor does

AI Token Monitor reads the session logs that tools such as Claude Code, Codex, OpenCode, Gemini CLI, GJC, Grok, Kiro, OmO, Pi and Hermes Agent already write, then prices each token with per-model rates, cache reads included. Today's spend sits next to the clock, and a dashboard adds daily charts, an activity heatmap, model breakdowns and cache efficiency. Live five-hour session and weekly plan bars can warn you before limits, with alerts to Discord, Slack or Telegram.

It works offline by default. The README says the app makes zero network requests unless you opt in to the leaderboard, chat or webhooks, and that only aggregated counts are shared then, never code or conversation content. A Rust engine deduplicates log entries and re-aggregates whenever a file changes. Extras include badges for a GitHub README, a monthly recap card and a receipt-style usage summary. Downloads are a macOS Apple Silicon dmg and a Windows 10+ installer.

Key features

• Reads local session logs from ten tools, including Hermes Agent
• Per-model pricing with cache reads, shown in the menu bar or system tray
• Daily and monthly charts, activity heatmap and model breakdown
• Plan-limit bars with Discord, Slack and Telegram alerts
• Offline by default, with opt-in leaderboard, chat and webhooks

When to use it

• Seeing today's spend across Hermes Agent and other tools in one place
• Getting a webhook alert before a plan limit is reached
• Comparing model cost and cache efficiency over a month

Who it is for: Developers who use several AI coding agents and want a local view of what they cost.

How it fits with Hermes Agent

Hermes Agent is one of ten log sources it reads, so this is a general usage monitor rather than a Hermes-specific tool.

Requirements: macOS on Apple Silicon (dmg) or Windows 10+ (installer; WebView2 is auto-installed).

FAQ