Stats For Nerds

v1.0.0可观测与分析

opencode-stats-for-nerds

OpenCode TUI plugin — token usage, speed, and cost stats in the sidebar

维护活跃

#opencode #opencode-plugin #stats #tokens #tui #sidebar

GitHub 仓库 npm 页面官方主页

GitHub 星标

月装机量

近 7 天 10

综合评分SCORE

30.5

生态多维模型

快速安装与配置

opencode.json

写入当前项目的 opencode.json，只对这个仓库生效。

opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "plugin": ["opencode-stats-for-nerds@1.0.0"]
}

写入 ~/.config/opencode/opencode.json，对所有项目生效。

~/.config/opencode/opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "plugin": ["opencode-stats-for-nerds@1.0.0"]
}

opencode 启动时会通过内嵌运行时自动加载 npm 依赖并缓存至本地目录，无需手动在全局环境执行安装。

README.md

查看原始文件→

OpenCode TUI plugin that shows token usage, context window, generation speed, cost, and file changes in the sidebar after each response.

Install

opencode plugin opencode-stats-for-nerds

Restart OpenCode. A Stats for Nerds panel appears in the right sidebar.

What it shows

▼ Stats for Nerds
  Tokens       13.4k in · 1.2k out · 2.0k thinking
  Cached       3.8k read · 2.7k write
  Context      23.2k / 200k (11%)
  ────────────────
  Cost         $0.0070
  Gen Time     2m 14s
  TTFT         1.35s
  Speed        42.3 tok/s
  ────────────────
  Changes      +234 -56 · 5 files
  Model        claude-sonnet-4

Click the header to collapse/expand.

Stats reference

Stat	Scope	Description
Tokens	Cumulative	Input, output, and thinking tokens across all turns
Cached	Cumulative	Prompt cache hits (read = reused, write = newly cached)
Context	Current	Context window used vs model limit, with percentage
Cost	Cumulative	Total session cost at provider rates
Gen Time	Cumulative	Total time the model spent generating
Think Time	Cumulative	Total time the model spent reasoning (off by default)
TTFT	Last response	Time-to-first-token
Speed	Last response	Output tokens per second
Activity	Cumulative	Tool-call steps and invocations (off by default)
Changes	Session	File additions, deletions, and count
Model	Last response	Model ID

Configuration

Disable any stat via tui.json:

{
  "$schema": "https://opencode.ai/tui.json",
  "plugin": [
    ["opencode-stats-for-nerds", {
      "show": {
        "cache": false,
        "activity": true,
        "model": false
      }
    }]
  ]
}

All options default to true except thinkTime and activity.

Key	Default	Controls
`tokens`	`true`	Input/output/thinking row
`cache`	`true`	Cache read/write row
`context`	`true`	Context window used / limit (%)
`cost`	`true`	Session cost
`genTime`	`true`	Total generation time
`thinkTime`	`false`	Total reasoning time
`ttft`	`true`	Time-to-first-token
`speed`	`true`	Tokens per second
`activity`	`false`	Steps and tool calls
`changes`	`true`	File additions/deletions
`model`	`true`	Model name

Requirements

OpenCode v1.17+

License

MIT