AI Agent Hub
Back to plugins
🖥️

dsh-peak-balance

Client Updated 2026.09.16

Run the following command in DeepSeek Harness:

dsh plugin install VviLliAm-qwq/dsh-peak-balance

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

Run dsh plugin install VviLliAm-qwq/dsh-peak-balance in the DeepSeek Harness terminal to install this plugin; the source repository is available at https://github.com/VviLliAm-qwq/dsh-peak-balance . Restart the TUI after installation to activate it.

About this plugin

DeepSeek applies peak-hour pricing during specific weekday windows, yet most clients give you no at-a-glance view of the current billing phase, your remaining balance, or what the turn you just finished actually cost. dsh-peak-balance embeds all three into the dsh-TUI status line: a live peak/off-peak clock with countdown, a provider-aware quota readout that follows the focused conversation, and a per-turn cost figure that ticks in real time while the model is generating and freezes the moment the turn closes.

Beyond the status line, the plugin ships a full-screen /hist token-history grid rendered in the familiar GitHub-contribution style. Hover any day for an input/cache-read/output breakdown, estimated cost, and cache-hit rate; arrow keys walk the grid, and single-key shortcuts cycle metric, time span, and subagent inclusion. During peak hours the status line can transform into a pulsing bordered frame in any of seven colors. All figures are sourced from this machine's session logs: token counts are the provider's own reported values, while costs are rate-card estimates and are labelled as such.

It is built for developers and teams who call the DeepSeek API heavily each day and want billing windows, live balance, and per-turn spend visible in their workflow, plus a glanceable grid that maps weeks of usage, model mix, and cache efficiency at a single command.

Use Cases

  • Watch the live peak/off-peak phase, account quota, and per-turn cost directly in the dsh-TUI status line
  • Open /hist to scan weeks of token usage, estimated cost, and cache-hit rate in a contribution-style grid
  • Enable a pulsing status-frame alert during peak windows so you pause before paying the higher rate

Best For

  • Solo developers who call the DeepSeek API heavily and want cost visible in their workflow
  • Teams juggling multiple projects who need to track subscription windows or relay credit pools
  • Tech leads who want per-day, per-model usage review and custom rate definitions for unrated models