AI Agent Hub
Back to plugins
🖥️

dsh-token-usage

Client Updated 2026.08.17

Run the following command in DeepSeek Harness:

dsh plugin install KeFan-J/dsh-token-usage

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install KeFan-J/dsh-token-usage in your terminal; the source lives at https://github.com/KeFan-J/dsh-token-usage. After installation, restart DSH and open the Token Usage section under Settings.

About this plugin

Every model call in DSH quietly burns tokens, yet users have had no built-in way to see the numbers without digging through raw session logs. dsh-token-usage turns that invisible ledger into a single dashboard under Settings with no extra configuration, ready the moment you open it.

It does two things at once: it subscribes to the live event stream of every active session and, on startup, replays the full history of persisted session logs. That means the very first render already includes usage from before you installed the plugin, and nothing is lost across restarts. The dashboard surfaces daily, weekly, and monthly totals broken down by input, output, and cache hits; computes thinking duration from reasoning-delta fragments; renders a 14-day stacked bar chart with hover tooltips; ranks models by weekly token share; groups usage by workspace; and estimates cost using DeepSeek official peak and off-peak pricing, which you can fetch with one click or edit by hand.

It is built for power developers who push DSH for hours a day, tech leads juggling multiple workspaces, and anyone who wants a single, honest answer to how much this month actually cost. All data stays on your machine; the only outbound request is the optional price-sync button that reads DeepSeek public pricing page.

Use Cases

  • Reviewing monthly token spend and cost against the actual bill
  • Comparing token share and peak/off-peak cost across models in one project
  • Spotting usage spikes at a glance from the 14-day stacked chart

Best For

  • AI developers who depend on DSH for daily work
  • Tech leads tracking token budgets across multiple workspaces
  • Individuals who want a clear picture of DeepSeek API costs