AI Agent Hub
Back to plugins
🖥️

dsh-balance

Client Updated 2026.08.24

Run the following command in DeepSeek Harness:

dsh plugin install YYfather/dsh-balance

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

Run dsh plugin install YYfather/dsh-balance in your DeepSeek Harness terminal to install; the full source is available at https://github.com/YYfather/dsh-balance

About this plugin

Chatting with DeepSeek or MiMo is convenient, but tracking what each session truly costs is not. dsh-balance surfaces account balance and per-request spend right in the composer dock, so you can see remaining credits and current burn rate at a glance without parsing invoices or converting tokens to dollars by hand.

Pricing here is granular, not a blanket rate: every request is matched to its peak or off-peak tier by timestamp, priced according to the exact model it invoked, and its usage samples are de-duplicated within a step per official projection rules. Four headline numbers -- session total, spend since plugin enable, cost of the latest request, and the previous session's total -- are all computed per-request. Built-in tables ship with DeepSeek V4 Flash, V4 Pro, and Xiaomi MiMo V2.5 at official dual-tier rates, and you can edit any model-plus-time-slot price or adjust the USD-to-CNY exchange rate from the click panel. Once active spend hits the alert line (default 1 CNY), the status bar flashes an orange badge to flag the overrun.

If you hop between MiMo for quick chat and DeepSeek for long generations, or run multiple models in parallel to compare output quality, dsh-balance gives you the one number that is usually missing: real cost. Keys are resolved on the Host side only, and plugin routes deny non-local origins by default, so the security model stays simple for everyday use.

Use Cases

  • Check DeepSeek account balance mid-conversation without leaving the chat
  • Compare peak vs. off-peak costs for the same model side by side
  • Get an orange badge alert when active spend crosses your threshold

Best For

  • Developers who call the DeepSeek API frequently
  • Users running DeepSeek and MiMo models in parallel
  • Individuals or teams that need to cap monthly token spend