Results (Japan, fetched 2026-10-13)
1. Trending now (JP)
Nothing in today's Japan top trending list matches AI. The feed returned 10 trending items, and none matched my filter. The filter terms were ai, claude, agent, chatgpt, gemini, openai, 生成AI, 人工知能, エージェント and コーディング. That is a valid result. It doesn't mean nobody is searching for AI. It means AI isn't in today's top 10. So there's no "ride today's spike" angle for AI coding agents. Evergreen search demand is the better target.
2. Autocomplete around "claude code"
These are what people type. They aren't a ranking and carry no traffic numbers.
| Intent cluster | Japanese queries | English queries |
|---|---|---|
| How-to / beginner | claude code 使い方, 使い方 初心者, 始め方, できること, とは | install, docs |
| Pricing / free | claude code 料金, 無料, クロードコード プラン | pricing |
| Install / setup | インストール, 使い方 windows, 使い方 mac, 使い方 vscode | install, download, cli, desktop |
| Comparison / confusion | claude codeとclaudeの違い | (none) |
| Mobile / app / login | クロードコード スマホ, アプリ, ログイン | desktop |
| Learning material | claude code 使い方 本, 使い方 日本語 | github, skills |
Katakana (クロードコード) is a real query form. Many Japanese users search it instead of the Latin spelling.
Keywords worth writing about
- claude code 使い方 初心者: the strongest beginner entry point. It also has long-tail variants (vscode, windows, mac, デスクトップ), so one pillar article can be followed by OS- and editor-specific posts.
- claude code 料金 / 無料 / プラン: pricing and free-tier questions show up in both the Latin and katakana forms. These readers are ready to decide, so this is a good place for a comparison table.
- claude codeとclaudeの違い: a specific confusion query with little to compete against. It suits a short explainer.
- クロードコードとは / できること: the "what is it" intent in katakana. Use the katakana spelling in the title and headings.
- クロードコード スマホ / アプリ: people want to know if it works on mobile. This could be a quick, differentiated post.
- claude code インストール: a practical setup guide covering Windows and Mac.
- claude code 使い方 本: signals demand for structured learning, such as a guide or ebook.
Suggestion: write one Japanese pillar article covering とは → インストール → 使い方 → 料金. Then add spin-offs for the OS/editor variants and the "違い" question. Use both spellings (Claude Code and クロードコード) in titles and body text.
Limits: Trending data is the top 10 only, with approximate traffic buckets. Autocomplete is not a ranking and has no volume numbers. I can't tell you which of these keywords has the most searches. For that you'd need Google Keyword Planner or Search Console.
Real run recorded with claude-code / claude-sonnet-5-5. Output is shown verbatim, unmodified.
它做什么
回答"X 相关的人们现在在搜什么"，数据是刚刚现取的，不靠模型记忆。它结合两个公开来源：Google Trends 今日热搜（每个国家前 10 条左右，带大致搜索量级和新闻链接）和 Google 搜索联想词（人们围绕某个词实际输入的内容）。
工作方式
- 从你的问题里提取 1 到 3 个核心关键词，并确定国家和语言。
- 运行自带的 Python 脚本（仅标准库），读取两个公开地址。
- 按主题过滤热搜，并把联想词按意图分组（教程、价格、对比、安装等）。
- 报告相关词；没有匹配或请求失败时会如实说明。
适合场景
内容与 SEO 选题、主题调研、查看某个话题在某国是否正在热搜。
局限
每个国家只开放前 10 条左右的热搜；搜索量只是大致区间；联想词不是排名，也没有搜索量。拿不到历史曲线和精确搜索量。
中风险：包含 1 个 Python 脚本，会向 Google 的两个公开地址（trends.google.com 的 RSS 和 suggestqueries.google.com）发起网络请求。不需要账号或密钥，不读取本地文件，也不写入任何文件。你的关键词会发送给 Google。这是非官方的公开接口，可能被限流或变更。代理需要有运行 `python3` 的权限。