cc-speed 0.1.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,86 @@
+[English](#) | [简体中文](./README.zh-CN.md)
+
+# cc-speed
+
+Analyze Claude Code token output speed from local JSONL logs.
+
+## What it does
+
+`cc-speed` reads your local Claude Code conversation logs (`~/.claude/projects/**/*.jsonl`) and calculates token output speed (tokens/sec) by measuring the time between user and assistant messages.
+
+It provides:
+- Per-model statistics (median, avg, p10, p90)
+- Daily trend sparklines in terminal
+- Interactive HTML charts (scatter plot + histogram) with dark theme
+
+## Install
+
+```bash
+npm install -g cc-speed
+```
+
+Or run directly:
+
+```bash
+npx cc-speed
+```
+
+## Usage
+
+```bash
+# Default: terminal table for the last 7 days
+cc-speed
+
+# Custom time range
+cc-speed --days 30
+
+# Generate HTML chart and open in browser
+cc-speed --chart
+
+# JSON output (for piping)
+cc-speed --json
+```
+
+### Options
+
+| Flag | Description | Default |
+|------|-------------|---------|
+| `-d, --days <n>` | Number of days to analyze | `7` |
+| `-c, --chart` | Generate HTML chart and open in browser | - |
+| `-j, --json` | Output as JSON | - |
+
+## Claude Code Slash Command
+
+You can use `cc-speed` as a `/speed` command inside Claude Code:
+
+```bash
+# One-line install
+bash <(curl -fsSL https://raw.githubusercontent.com/bryant24hao/cc-speed/main/install-skill.sh)
+```
+
+Or manually create `~/.claude/skills/cc-speed/SKILL.md`:
+
+```markdown
+---
+name: speed
+description: Analyze Claude Code token output speed
+---
+
+Run `npx cc-speed --chart` to analyze token output speed and generate charts.
+```
+
+Then type `/speed` in Claude Code to trigger it.
+
+## How it works
+
+1. Scans `~/.claude/projects/**/*.jsonl` for recent conversation logs
+2. Pairs each `assistant` message with its preceding `user` message
+3. Calculates `output_tokens / (assistant_timestamp - user_timestamp)`
+4. Filters: `output_tokens > 10`, duration between 0.5s and 300s
+5. Groups by model, computes statistics and trends
+
+> **Note**: The speed is an estimate based on timestamp deltas. It includes network latency and any processing overhead, so actual generation speed may be higher.
+
+## License
+
+MIT

package/README.zh-CN.md

@@ -0,0 +1,86 @@
+[English](./README.md) | [简体中文](#)
+
+# cc-speed
+
+分析 Claude Code 的 token 输出速度。
+
+## 功能介绍
+
+`cc-speed` 读取本地 Claude Code 的对话日志（`~/.claude/projects/**/*.jsonl`），通过测量 user 和 assistant 消息之间的时间差来计算 token 输出速度（tokens/sec）。
+
+提供：
+- 按模型分组的速度统计（中位数、平均值、P10、P90）
+- 终端中的每日趋势 sparkline 图
+- 交互式 HTML 图表（散点图 + 直方图），暗色主题
+
+## 安装
+
+```bash
+npm install -g cc-speed
+```
+
+或直接运行：
+
+```bash
+npx cc-speed
+```
+
+## 使用方法
+
+```bash
+# 默认：终端表格展示最近 7 天
+cc-speed
+
+# 自定义时间范围
+cc-speed --days 30
+
+# 生成 HTML 图表并在浏览器打开
+cc-speed --chart
+
+# JSON 输出（方便管道处理）
+cc-speed --json
+```
+
+### 参数
+
+| 参数 | 说明 | 默认值 |
+|------|------|--------|
+| `-d, --days <n>` | 分析天数 | `7` |
+| `-c, --chart` | 生成 HTML 图表并在浏览器打开 | - |
+| `-j, --json` | 输出 JSON 格式 | - |
+
+## Claude Code Slash Command
+
+可以在 Claude Code 中通过 `/speed` 命令使用：
+
+```bash
+# 一行命令安装
+bash <(curl -fsSL https://raw.githubusercontent.com/bryant24hao/cc-speed/main/install-skill.sh)
+```
+
+或手动创建 `~/.claude/skills/cc-speed/SKILL.md`：
+
+```markdown
+---
+name: speed
+description: 分析 Claude Code 的 token 输出速度
+---
+
+运行 `npx cc-speed --chart` 分析 token 输出速度并生成图表。
+```
+
+然后在 Claude Code 中输入 `/speed` 触发。
+
+## 工作原理
+
+1. 扫描 `~/.claude/projects/**/*.jsonl` 中的对话日志
+2. 将每条 `assistant` 消息与其前一条 `user` 消息配对
+3. 计算 `output_tokens / (assistant 时间戳 - user 时间戳)`
+4. 过滤条件：`output_tokens > 10`，时长在 0.5s 到 300s 之间
+5. 按模型分组，计算统计数据和趋势
+
+> **注意**：速度是基于时间戳差值的估算，包含了网络延迟和处理开销，实际生成速度可能更高。
+
+## 许可证
+
+MIT

package/dist/index.js

@@ -0,0 +1,449 @@
+#!/usr/bin/env node
+
+// src/index.ts
+import { program } from "commander";
+import ora from "ora";
+
+// src/collector.ts
+import { readFileSync, readdirSync, statSync } from "fs";
+import { join } from "path";
+import { homedir } from "os";
+function getAllJsonlFiles(dir, cutoffMs) {
+  const results = [];
+  try {
+    for (const entry of readdirSync(dir, { withFileTypes: true })) {
+      const fullPath = join(dir, entry.name);
+      if (entry.isDirectory()) {
+        results.push(...getAllJsonlFiles(fullPath, cutoffMs));
+      } else if (entry.name.endsWith(".jsonl")) {
+        try {
+          const stat = statSync(fullPath);
+          if (stat.mtimeMs > cutoffMs) {
+            results.push(fullPath);
+          }
+        } catch {
+        }
+      }
+    }
+  } catch {
+  }
+  return results;
+}
+function parseJsonlFile(filePath) {
+  const messages = [];
+  try {
+    const lines = readFileSync(filePath, "utf-8").split("\n").filter(Boolean);
+    for (const line of lines) {
+      try {
+        messages.push(JSON.parse(line));
+      } catch {
+      }
+    }
+  } catch {
+  }
+  return messages;
+}
+function normalizeModel(msg) {
+  const raw = msg.message?.model || msg.model || "unknown";
+  return raw.replace("claude-", "").replace(/-\d{8}$/, "");
+}
+function collectSpeedData(days) {
+  const projectsDir = join(homedir(), ".claude", "projects");
+  const cutoff = /* @__PURE__ */ new Date();
+  cutoff.setDate(cutoff.getDate() - days);
+  const cutoffMs = cutoff.getTime();
+  const files = getAllJsonlFiles(projectsDir, cutoffMs);
+  const dataPoints = [];
+  for (const file of files) {
+    const messages = parseJsonlFile(file);
+    for (let i = 0; i < messages.length; i++) {
+      const msg = messages[i];
+      if (msg.type !== "assistant" || !msg.message?.usage?.output_tokens) continue;
+      if (!msg.timestamp) continue;
+      const ts = new Date(msg.timestamp);
+      if (ts.getTime() < cutoffMs) continue;
+      const outputTokens = msg.message.usage.output_tokens;
+      const model = normalizeModel(msg);
+      let userTs = null;
+      for (let j = i - 1; j >= 0; j--) {
+        if (messages[j].type === "user" && messages[j].timestamp) {
+          userTs = new Date(messages[j].timestamp);
+          break;
+        }
+      }
+      if (userTs && outputTokens > 10) {
+        const durationSec = (ts.getTime() - userTs.getTime()) / 1e3;
+        if (durationSec > 0.5 && durationSec < 300) {
+          dataPoints.push({
+            timestamp: ts.toISOString(),
+            outputTokens,
+            model,
+            durationSec: Math.round(durationSec * 10) / 10,
+            tokensPerSec: Math.round(outputTokens / durationSec * 10) / 10
+          });
+        }
+      }
+    }
+  }
+  dataPoints.sort((a, b) => new Date(a.timestamp).getTime() - new Date(b.timestamp).getTime());
+  return dataPoints;
+}
+
+// src/analyzer.ts
+function percentile(sorted, p) {
+  const idx = Math.floor(sorted.length * p);
+  return sorted[Math.min(idx, sorted.length - 1)];
+}
+function median(sorted) {
+  const mid = Math.floor(sorted.length / 2);
+  if (sorted.length % 2 === 0) {
+    return Math.round((sorted[mid - 1] + sorted[mid]) / 2 * 10) / 10;
+  }
+  return sorted[mid];
+}
+function analyzeSpeed(data, days = 7) {
+  const byModel = /* @__PURE__ */ new Map();
+  for (const dp of data) {
+    const arr = byModel.get(dp.model) || [];
+    arr.push(dp);
+    byModel.set(dp.model, arr);
+  }
+  const modelStats = [];
+  for (const [model, points] of byModel) {
+    const speeds = points.map((p) => p.tokensPerSec).sort((a, b) => a - b);
+    modelStats.push({
+      model,
+      count: speeds.length,
+      median: median(speeds),
+      avg: Math.round(speeds.reduce((a, b) => a + b, 0) / speeds.length * 10) / 10,
+      p10: percentile(speeds, 0.1),
+      p90: percentile(speeds, 0.9),
+      min: speeds[0],
+      max: speeds[speeds.length - 1]
+    });
+  }
+  modelStats.sort((a, b) => a.model.localeCompare(b.model));
+  const dailyMap = /* @__PURE__ */ new Map();
+  for (const dp of data) {
+    const date = dp.timestamp.slice(0, 10);
+    const key = `${date}|${dp.model}`;
+    const arr = dailyMap.get(key) || [];
+    arr.push(dp);
+    dailyMap.set(key, arr);
+  }
+  const dailyTrends = [];
+  for (const [key, points] of dailyMap) {
+    const [date, model] = key.split("|");
+    const speeds = points.map((p) => p.tokensPerSec).sort((a, b) => a - b);
+    dailyTrends.push({
+      date,
+      model,
+      median: median(speeds),
+      count: speeds.length
+    });
+  }
+  dailyTrends.sort((a, b) => a.date.localeCompare(b.date) || a.model.localeCompare(b.model));
+  return {
+    modelStats,
+    dailyTrends,
+    totalSamples: data.length,
+    days
+  };
+}
+
+// src/terminal.ts
+import chalk from "chalk";
+import Table from "cli-table3";
+var MODEL_COLORS = {
+  opus: chalk.hex("#f78166"),
+  sonnet: chalk.hex("#7ee787"),
+  haiku: chalk.hex("#79c0ff")
+};
+function colorForModel(model) {
+  for (const [key, colorFn] of Object.entries(MODEL_COLORS)) {
+    if (model.toLowerCase().includes(key)) return colorFn;
+  }
+  return chalk.white;
+}
+function sparkline(values) {
+  if (values.length === 0) return "";
+  const bars = "\u2581\u2582\u2583\u2584\u2585\u2586\u2587\u2588";
+  const min = Math.min(...values);
+  const max = Math.max(...values);
+  const range = max - min || 1;
+  return values.map((v) => {
+    const idx = Math.round((v - min) / range * (bars.length - 1));
+    return bars[idx];
+  }).join("");
+}
+function renderTerminal(report) {
+  if (report.totalSamples === 0) {
+    console.log(chalk.yellow("\nNo data found. Make sure you have Claude Code conversation logs in ~/.claude/projects/\n"));
+    return;
+  }
+  console.log(
+    chalk.bold(`
+  Claude Code Output Speed (last ${report.days} days, ${report.totalSamples} samples)
+`)
+  );
+  const table = new Table({
+    head: ["Model", "Samples", "Median", "Avg", "P10", "P90", "Min", "Max"].map(
+      (h) => chalk.gray(h)
+    ),
+    style: { head: [], border: ["gray"] },
+    chars: {
+      top: "\u2500",
+      "top-mid": "\u252C",
+      "top-left": "\u250C",
+      "top-right": "\u2510",
+      bottom: "\u2500",
+      "bottom-mid": "\u2534",
+      "bottom-left": "\u2514",
+      "bottom-right": "\u2518",
+      left: "\u2502",
+      "left-mid": "\u251C",
+      mid: "\u2500",
+      "mid-mid": "\u253C",
+      right: "\u2502",
+      "right-mid": "\u2524",
+      middle: "\u2502"
+    }
+  });
+  for (const stat of report.modelStats) {
+    const color = colorForModel(stat.model);
+    table.push([
+      color(stat.model),
+      String(stat.count),
+      chalk.bold(stat.median + " tok/s"),
+      stat.avg + " tok/s",
+      stat.p10 + "",
+      stat.p90 + "",
+      stat.min + "",
+      stat.max + ""
+    ]);
+  }
+  console.log(table.toString());
+  if (report.dailyTrends.length > 0) {
+    console.log(chalk.bold("\n  Daily Trend (median tok/s)\n"));
+    const models = [...new Set(report.dailyTrends.map((t) => t.model))];
+    const dates = [...new Set(report.dailyTrends.map((t) => t.date))].sort();
+    for (const model of models) {
+      const color = colorForModel(model);
+      const dailyValues = dates.map((date) => {
+        const found = report.dailyTrends.find((t) => t.date === date && t.model === model);
+        return found ? found.median : 0;
+      });
+      const nonZero = dailyValues.filter((v) => v > 0);
+      if (nonZero.length === 0) continue;
+      const spark = sparkline(nonZero);
+      const latest = nonZero[nonZero.length - 1];
+      console.log(`  ${color(model.padEnd(16))} ${spark}  ${chalk.bold(latest + "")} tok/s`);
+    }
+  }
+  console.log();
+}
+
+// src/chart.ts
+import { writeFileSync } from "fs";
+import { join as join2 } from "path";
+import { tmpdir } from "os";
+function generateChart(data, report) {
+  let statCards = "";
+  for (const stat of report.modelStats) {
+    const cls = stat.model.includes("opus") ? "opus" : stat.model.includes("haiku") ? "haiku" : "sonnet";
+    statCards += `<div class="stat-card">`;
+    statCards += `<div class="stat-label">${stat.model}</div>`;
+    statCards += `<div class="stat-value ${cls}">${stat.median} <span style="font-size:14px;color:#8b949e">tok/s</span></div>`;
+    statCards += `<div class="stat-label">median &middot; avg ${stat.avg} &middot; p10 ${stat.p10} &middot; p90 ${stat.p90} &middot; n=${stat.count}</div>`;
+    statCards += `</div>`;
+  }
+  const html = `<!DOCTYPE html>
+<html lang="en"><head><meta charset="utf-8">
+<title>Claude Code Token Output Speed</title>
+<script src="https://cdn.jsdelivr.net/npm/chart.js@4.4.7/dist/chart.umd.min.js"></script>
+<script src="https://cdn.jsdelivr.net/npm/chartjs-adapter-date-fns@3.0.0/dist/chartjs-adapter-date-fns.bundle.min.js"></script>
+<style>
+body { font-family: -apple-system, BlinkMacSystemFont, sans-serif; background: #0d1117; color: #e6edf3; padding: 24px; margin: 0; }
+h1 { font-size: 20px; font-weight: 600; margin-bottom: 4px; }
+.subtitle { color: #8b949e; font-size: 14px; margin-bottom: 24px; }
+.chart-container { background: #161b22; border-radius: 12px; padding: 20px; margin-bottom: 20px; border: 1px solid #30363d; }
+.stats { display: flex; gap: 16px; margin-bottom: 20px; flex-wrap: wrap; }
+.stat-card { background: #161b22; border: 1px solid #30363d; border-radius: 8px; padding: 16px 20px; min-width: 140px; }
+.stat-label { color: #8b949e; font-size: 12px; margin-bottom: 4px; }
+.stat-value { font-size: 24px; font-weight: 700; }
+.opus { color: #f78166; }
+.sonnet { color: #7ee787; }
+.haiku { color: #79c0ff; }
+canvas { max-height: 400px; }
+.lang-toggle { position: fixed; top: 16px; right: 24px; background: #21262d; border: 1px solid #30363d; border-radius: 6px; padding: 4px; display: flex; gap: 2px; z-index: 100; }
+.lang-toggle button { background: none; border: none; color: #8b949e; padding: 4px 10px; border-radius: 4px; cursor: pointer; font-size: 13px; }
+.lang-toggle button.active { background: #30363d; color: #e6edf3; }
+</style></head><body>
+<div class="lang-toggle">
+  <button id="btn-en" class="active" onclick="setLang('en')">EN</button>
+  <button id="btn-zh" onclick="setLang('zh')">\u4E2D\u6587</button>
+</div>
+<h1 id="title">Claude Code Output Speed (tokens/sec)</h1>
+<div class="subtitle" id="subtitle">Last ${report.days} days &middot; ${report.totalSamples} samples &middot; based on user&rarr;assistant timestamp delta &middot; output_tokens &gt; 10</div>
+<div class="stats">${statCards}</div>
+<div class="chart-container"><canvas id="speedChart"></canvas></div>
+<div class="chart-container"><canvas id="histChart"></canvas></div>
+<script>
+var i18n = {
+  en: {
+    title: "Claude Code Output Speed (tokens/sec)",
+    subtitle: "Last ${report.days} days \\u00b7 ${report.totalSamples} samples \\u00b7 based on user\\u2192assistant timestamp delta \\u00b7 output_tokens > 10",
+    scatterTitle: "Output Speed Over Time",
+    histTitle: "Speed Distribution (tokens/sec)",
+    ySpeed: "tokens/sec",
+    yCount: "count"
+  },
+  zh: {
+    title: "Claude Code \\u8f93\\u51fa\\u901f\\u5ea6 (tokens/sec)",
+    subtitle: "\\u6700\\u8fd1 ${report.days} \\u5929 \\u00b7 ${report.totalSamples} \\u4e2a\\u6837\\u672c \\u00b7 \\u57fa\\u4e8e user\\u2192assistant \\u65f6\\u95f4\\u5dee\\u4f30\\u7b97 \\u00b7 \\u4ec5\\u7edf\\u8ba1 output_tokens > 10",
+    scatterTitle: "\\u8f93\\u51fa\\u901f\\u5ea6\\u968f\\u65f6\\u95f4\\u53d8\\u5316",
+    histTitle: "\\u901f\\u5ea6\\u5206\\u5e03 (tokens/sec)",
+    ySpeed: "tokens/sec",
+    yCount: "\\u6570\\u91cf"
+  }
+};
+var currentLang = "en";
+
+function setLang(lang) {
+  currentLang = lang;
+  document.getElementById("btn-en").className = lang === "en" ? "active" : "";
+  document.getElementById("btn-zh").className = lang === "zh" ? "active" : "";
+  document.getElementById("title").textContent = i18n[lang].title;
+  document.getElementById("subtitle").textContent = i18n[lang].subtitle;
+  if (window.scatterChart) {
+    window.scatterChart.options.plugins.title.text = i18n[lang].scatterTitle;
+    window.scatterChart.options.scales.y.title.text = i18n[lang].ySpeed;
+    window.scatterChart.update();
+  }
+  if (window.histChart) {
+    window.histChart.options.plugins.title.text = i18n[lang].histTitle;
+    window.histChart.options.scales.y.title.text = i18n[lang].yCount;
+    window.histChart.options.scales.x.title.text = i18n[lang].ySpeed;
+    window.histChart.update();
+  }
+}
+
+const data = ${JSON.stringify(data)};
+const models = [...new Set(data.map(d => d.model))];
+const colors = {
+  'opus-4-6': '#f78166',
+  'sonnet-4-5': '#7ee787',
+  'haiku-4-5': '#79c0ff',
+};
+function getColor(model) {
+  if (colors[model]) return colors[model];
+  if (model.includes('opus')) return '#f78166';
+  if (model.includes('haiku')) return '#79c0ff';
+  return '#7ee787';
+}
+
+window.scatterChart = new Chart(document.getElementById("speedChart"), {
+  type: "scatter",
+  data: {
+    datasets: models.map(model => ({
+      label: model,
+      data: data.filter(d => d.model === model).map(d => ({ x: new Date(d.timestamp), y: d.tokensPerSec })),
+      backgroundColor: getColor(model) + "99",
+      borderColor: getColor(model),
+      pointRadius: 4,
+      pointHoverRadius: 6,
+    })),
+  },
+  options: {
+    responsive: true,
+    plugins: {
+      title: { display: true, text: i18n[currentLang].scatterTitle, color: "#e6edf3" },
+      legend: { labels: { color: "#e6edf3" } },
+      tooltip: {
+        callbacks: {
+          label: (ctx) => {
+            const pts = data.filter(dd => dd.model === ctx.dataset.label);
+            const d = pts[ctx.dataIndex];
+            if (!d) return "";
+            return d.tokensPerSec + " tok/s \\u00b7 " + d.outputTokens + " tokens \\u00b7 " + d.durationSec + "s";
+          }
+        }
+      }
+    },
+    scales: {
+      x: { type: "time", time: { unit: "day" }, ticks: { color: "#8b949e" }, grid: { color: "#21262d" } },
+      y: { title: { display: true, text: i18n[currentLang].ySpeed, color: "#8b949e" }, ticks: { color: "#8b949e" }, grid: { color: "#21262d" } }
+    }
+  }
+});
+
+var bucketSize = 5;
+var maxSpeed = Math.ceil(Math.max(...data.map(d => d.tokensPerSec)) / bucketSize) * bucketSize;
+var buckets = Array.from({ length: maxSpeed / bucketSize + 1 }, (_, i) => i * bucketSize);
+var histDataMap = {};
+models.forEach(function(model) {
+  histDataMap[model] = new Array(buckets.length).fill(0);
+  data.filter(dd => dd.model === model).forEach(function(d) {
+    var idx = Math.floor(d.tokensPerSec / bucketSize);
+    if (idx < histDataMap[model].length) histDataMap[model][idx]++;
+  });
+});
+
+window.histChart = new Chart(document.getElementById("histChart"), {
+  type: "bar",
+  data: {
+    labels: buckets.map(b => b + "-" + (b + bucketSize)),
+    datasets: models.map(model => ({
+      label: model,
+      data: histDataMap[model],
+      backgroundColor: getColor(model) + "99",
+      borderColor: getColor(model),
+      borderWidth: 1,
+    })),
+  },
+  options: {
+    responsive: true,
+    plugins: {
+      title: { display: true, text: i18n[currentLang].histTitle, color: "#e6edf3" },
+      legend: { labels: { color: "#e6edf3" } }
+    },
+    scales: {
+      x: { title: { display: true, text: i18n[currentLang].ySpeed, color: "#8b949e" }, ticks: { color: "#8b949e" }, grid: { color: "#21262d" } },
+      y: { title: { display: true, text: i18n[currentLang].yCount, color: "#8b949e" }, ticks: { color: "#8b949e" }, grid: { color: "#21262d" } }
+    }
+  }
+});
+</script></body></html>`;
+  const outPath = join2(tmpdir(), "claude-speed.html");
+  writeFileSync(outPath, html);
+  return outPath;
+}
+
+// src/index.ts
+import { exec } from "child_process";
+import { platform } from "os";
+program.name("cc-speed").description("Analyze Claude Code token output speed").version("0.1.0").option("-d, --days <number>", "Number of days to analyze", "7").option("-c, --chart", "Generate HTML chart and open in browser").option("-j, --json", "Output as JSON").action(async (opts) => {
+  const days = parseInt(opts.days, 10);
+  const spinner = ora("Scanning Claude Code logs...").start();
+  const data = collectSpeedData(days);
+  spinner.succeed(`Found ${data.length} data points from ${days} days`);
+  if (data.length === 0) {
+    console.log("\nNo data found. Make sure you have Claude Code conversation logs in ~/.claude/projects/\n");
+    return;
+  }
+  const report = analyzeSpeed(data, days);
+  if (opts.json) {
+    console.log(JSON.stringify(report, null, 2));
+    return;
+  }
+  if (opts.chart) {
+    const chartSpinner = ora("Generating chart...").start();
+    const chartPath = generateChart(data, report);
+    chartSpinner.succeed(`Chart saved to ${chartPath}`);
+    const openCmd = platform() === "darwin" ? "open" : platform() === "win32" ? "start" : "xdg-open";
+    exec(`${openCmd} "${chartPath}"`);
+    return;
+  }
+  renderTerminal(report);
+});
+program.parse();

package/install-skill.sh

@@ -0,0 +1,21 @@
+#!/usr/bin/env bash
+# Install cc-speed as a Claude Code /speed slash command
+
+set -e
+
+SKILL_DIR="$HOME/.claude/skills/cc-speed"
+SKILL_FILE="$SKILL_DIR/SKILL.md"
+
+mkdir -p "$SKILL_DIR"
+
+cat > "$SKILL_FILE" << 'EOF'
+---
+name: speed
+description: Analyze Claude Code token output speed / 分析 Claude Code 的 token 输出速度
+---
+
+Run `npx cc-speed --chart` to analyze token output speed and generate charts.
+EOF
+
+echo "Installed /speed command to $SKILL_FILE"
+echo "You can now use /speed in Claude Code."

package/package.json

@@ -0,0 +1,45 @@
+{
+  "name": "cc-speed",
+  "version": "0.1.0",
+  "description": "Analyze Claude Code token output speed from local JSONL logs",
+  "type": "module",
+  "bin": {
+    "cc-speed": "./dist/index.js"
+  },
+  "main": "./dist/index.js",
+  "files": [
+    "dist",
+    "install-skill.sh"
+  ],
+  "scripts": {
+    "build": "tsup",
+    "dev": "tsup --watch",
+    "test": "vitest run",
+    "prepublishOnly": "pnpm build"
+  },
+  "keywords": [
+    "claude",
+    "claude-code",
+    "token-speed",
+    "benchmark",
+    "cli"
+  ],
+  "author": "",
+  "license": "MIT",
+  "dependencies": {
+    "chalk": "^5.3.0",
+    "cli-table3": "^0.6.5",
+    "commander": "^12.1.0",
+    "glob": "^11.0.0",
+    "ora": "^8.1.0"
+  },
+  "devDependencies": {
+    "@types/node": "^22.0.0",
+    "tsup": "^8.3.0",
+    "typescript": "^5.7.0",
+    "vitest": "^3.0.0"
+  },
+  "pnpm": {
+    "onlyBuiltDependencies": ["esbuild"]
+  }
+}