deepseek-peak-pricing-hours

中文文档
0.2.5发布于 2026/11/03MITES Module零依赖

Pi extension showing DeepSeek API peak pricing hours (Beijing-time windows displayed in your local timezone)

原始使用文档
文档目录4 节
  1. Install
  2. How it works
  3. Commands
  4. License

deepseek-peak-pricing-hours

Visual indicator showing DeepSeek API peak pricing hours. Peak windows are defined in Beijing Time, but all times shown are converted to your local timezone.

Peak hours: 9:00–12:00 & 14:00–18:00 Beijing Time (UTC+8) — i.e. 01:00–04:00 & 06:00–10:00 UTC; off-peak tokens are billed at 50% of the peak rate.

On weekends (Saturday & Sunday, Beijing Time) the peak/off-peak division doesn't apply at all: all calls made during the day bill uniformly at the off-peak rate.

What it does

  • Renders a 24-hour bar below the editor showing peak vs off-peak hours
  • Shows a colored current-state diamond + label (◆ DeepSeek PEAK or ◇ DeepSeek Off-Peak)
  • Displays the next transition time in your local timezone
  • Auto-refreshes every 30 seconds and at session start
  • Shows the widget only when the selected model is a DeepSeek model (matched case-insensitively in its provider or model ID)
  • Automatically shows the widget when switching to a DeepSeek model and hides it when switching to any other model

Use /ds-peak to manually hide or show the widget while a DeepSeek model is selected. The command won't show the widget for other models; switching back to a DeepSeek model automatically shows it again.

Install

pi install npm:deepseek-peak-pricing-hours

Then restart pi, or run /reload.

Uninstall it using the same source you installed:

pi uninstall npm:deepseek-peak-pricing-hours

For a local checkout installed from this repository, run pi uninstall . from the repository root instead. Check pi list to see the configured source if you're unsure. Pi removes that package from its settings so it will no longer load in future sessions. The extension also clears its widget and refresh timer when the current session shuts down.

How it works

The bar uses block characters to visualize DeepSeek's two daily peak windows (weekdays only):

Peak WindowBeijing Time (UTC+8)Displayed in
Morning09:00 – 12:00Your local time
Afternoon14:00 – 18:00Your local time
Weekend (Sat & Sun)All day — flat off-peak, no peak windows applyYour local time
CharacterMeaning
█Peak hour (warning color)
░Off-peak hour (dim) — on weekends, every hour renders as off-peak
◆Current peak hour (bright warning/error color, bold)
◇Current off-peak hour (accent color)

On weekends the status line stays green (Off-Peak) for the whole day, and the next-transition caption points at Monday 09:00 Beijing time.

The bar collapses to 2 hours per character when narrow, and hides entirely in very tight spaces.

Commands

CommandWhat it does
/ds-peakToggle the indicator while a DeepSeek model is selected; unavailable for other models

License

MIT

版本与发布记录

deepseek-peak-pricing-hours 共发布 7 个版本,首次发布于 2026/07/10;过去 12 个月发布了 7 个版本。

版本发布日期距今
0.2.5latest2026/11/03今天
0.2.32026/08/242 个月前
0.2.22026/08/242 个月前
0.2.12026/08/232 个月前
0.2.02026/08/232 个月前
0.1.12026/07/103 个月前

关于 deepseek-peak-pricing-hours 的常见问题

deepseek-peak-pricing-hours 是什么?

Pi扩展显示DeepSeek API峰值定价时间(北京时间窗口显示在您的当地时区)

如何安装 deepseek-peak-pricing-hours?

在项目目录运行 npm install deepseek-peak-pricing-hours,或使用 pnpm add / yarn add / bun add。它还需要宿主项目提供 peer 依赖:@earendil-works/pi-tui(*)、@earendil-works/pi-coding-agent(*)。npm 7 及以上与 pnpm 8 及以上默认会自动安装必需的 peer 依赖,Yarn 需要手动安装。

deepseek-peak-pricing-hours 支持 TypeScript 吗?

该包的 package.json 没有声明类型入口。需要类型提示时,可以在 npm 上查找社区维护的 @types/deepseek-peak-pricing-hours 是否存在。

deepseek-peak-pricing-hours 支持 ESM 还是 CommonJS?

以 ES Module 形式发布,推荐使用 import 引入。在 CommonJS 项目中,Node.js 20.19+ / 22.12+ 可以直接 require() 不含顶层 await 的 ES Module,更早的版本需要改用动态 import()。

deepseek-peak-pricing-hours 可以商用吗?

deepseek-peak-pricing-hours 使用 MIT 许可。宽松许可:允许商业使用、修改和再分发,通常只需在分发时保留原版权与许可声明。以上为概括,不构成法律意见,请以许可原文为准。

deepseek-peak-pricing-hours 还在维护吗?

最新版本 0.2.5 发布于 2026/11/03(今天);至今共发布 7 个版本,其中过去 12 个月发布了 7 个;首次发布于 2026/07/10。发布频率只是参考信号:功能稳定的工具可能很少需要更新,建议结合源码仓库的 issue 和提交记录判断。