ikiwi26 downloadsArrange heading numbering, generate new headings, move images into folder, and insert blank lines. 整理标题序号,生成新标题,添加图片到文件夹内,插入空白行。
跳到中文。 Here is the English version:








Organize the current Markdown document with one command: renumber H1-H4 headings, and, according to settings, complete code block languages, merge blank lines, add blank lines at the start and end of blockquotes, and manage horizontal rules. There are also two standalone commands: one calls AI to regenerate headings based on the body text (then automatically organizes the document), and the other inserts a batch of blank lines after the current paragraph.
<note directory>/<image folder>/, rewrite links, and delete the original image (move semantics); supports ![[...]] and ; creates the folder automatically if it is missing1 / 1.1 / 1.1.1 / 1.1.1.1 (fixed Arabic numerals + . + space; all levels numbered; missing parent levels default to 1)> line at the start and end of each > quote block (do not add again if an empty > already exists; use the indentation of the block’s start/end)--- / *** / ___), and can insert --- between two H1 headings or between two H2 headingsclaude / deepseek / gemini / gpt / qwen / kimi| Setting | Default | Description |
|---|---|---|
| UI language | Chinese | Chinese / English; applies immediately to the settings page and command names (command IDs are unchanged, so hotkeys are kept) |
| Numbering mode | Single-level | Single-level: each level numbered independently; Multi-level: 1 / 1.1 / 1.1.1 / 1.1.1.1 (fixed Arabic numerals + . + space; per-level styles/delimiters do not apply) |
| Add space after delimiter | Off | When generating single-level numbering, add one space after the delimiter (automatically skipped for the “space” delimiter and Clear numbering) |
| Numbering system (per level) | Chinese numerals | 12 systems + Clear numbering (does not apply in multi-level mode) |
| Delimiter (per level) | 、 |
Used when generating; when recognizing existing numbering, compatible with 13 character delimiters (does not apply in multi-level mode) |
| Complete code block languages | Off | Only add language tags to fenced code blocks that lack a language |
| Merge blank lines | Off | Compress 3+ consecutive newlines into 1 blank line |
| Add blank lines around blockquotes | Off | Add one empty > line at the start and end of each > quote block (do not add if an empty > already exists) |
| Number of blank lines to insert | 20 | Number of lines inserted at once by the “Insert blank lines after this paragraph” command (1-100) |
| Auto-organize inserted images | Off | When the note is saved, move newly added images on this page to the image folder and rewrite links (move, not copy) |
| Image folder | picture |
Relative to the note’s directory; can be multi-level (assets/img); created automatically if it does not exist |
| Delete existing horizontal rules | Off | Delete --- / *** / ___; preserve frontmatter and setext underlines |
| Insert between two H1 headings | Off | Insert --- before every non-first H1 |
| Insert between two H2 headings | Off | Within each H1 group, insert --- before the group’s non-first H2; if there is no H1, process the whole document |
| Heading language | Follow document | Follow document (the body script is detected first — Chinese / English / Japanese / Korean / Russian / Arabic, Latin-script languages are split by common words — and the AI is required to use it; if the sample is too short the AI infers it) / Specify language (11 languages: Chinese, English, Japanese, Korean, Spanish, French, German, Portuguese, Russian, Italian, Arabic) |
| H1-H4 generated content | All on | emoji / discussion topic / one-sentence summary; at least one of the three; if all are off, the command reports an error and aborts |
| Heading generation prompt | Chinese default text | Generation requirements for “one-sentence summary”; read-only preview + edit (modal) + reset; if cleared, the default for the current UI language is used (so does a stored default that belongs to the other UI language) |
| Topic generation prompt | Chinese default text | Generation requirements for “discussion topic”; read-only preview + edit (modal) + reset; if cleared, the default for the current UI language is used (so does a stored default that belongs to the other UI language) |
| AI provider | Claude | Six providers: Claude / DeepSeek / Gemini / GPT / Qwen / Kimi; each has a built-in default model |
| API Key | Empty | Used only for “Regenerate headings”; displayed masked with type=password; can be cleared; saved in data.json |
| System | Limits |
|---|---|
| Chinese / Arabic / Eastern Arabic | Arabic and Eastern Arabic: 1..999 |
| Roman (uppercase/lowercase) | 1..3999; falls back to Arabic numerals beyond that |
| Latin (uppercase/lowercase) | Bijective base-26 (A..Z, AA..AB); parsing accepts only 1-2 letters |
| Circled numbers | ①-⑳, ㉑-㉟, ㊱-㊿; 51+ falls back to Arabic numerals |
| Hangul / circled Hangul | 14 each; falls back to Arabic numerals beyond that |
| Cyrillic | Russian alphabet skipping З / О (31 letters); falls back to Arabic numerals beyond that |
| Katakana | Iroha order, 47 characters; falls back to Arabic numerals beyond that |
用一个命令整理当前 Markdown 文档:重编号 H1-H4 标题,并按设置补全代码块语言、合并空行、为块引用补首尾空行、管理分割线;另有两个独立命令,分别用于调用 AI 按正文重新生成标题(随后自动整理)、在当前段落之后插入一批空白行。
<笔记目录>/<图片文件夹>/,改写链接并删除原图(移动语义);支持 ![[...]] 与 ,文件夹缺失时自动新建1 / 1.1 / 1.1.1 / 1.1.1.1 (固定阿拉伯数字 + . + 空格、全补号、缺父级按 1 兜底)> 引用段的首尾各补一行空 >(已有空 > 不重复添加,沿用段首/段尾缩进)--- / *** / ___),也可在两个 H1 之间或两个 H2 之间插入 ---claude / deepseek / gemini / gpt / qwen / kimi 六家服务商| 设置项 | 默认值 | 说明 |
|---|---|---|
| 界面语言 | 中文 | 中文 / English;设置页与命令名即时生效(命令 ID 不变,热键保留) |
| 编号模式 | 单级 | 单级:每级独立编号;多级:1 / 1.1 / 1.1.1 / 1.1.1.1(固定阿拉伯 + . + 空格,各级风格/分隔符不生效) |
| 分隔符后补空格 | 关 | 单级编号生成时在分隔符后加一个空格(「空格」分隔符、清除编号自动跳过) |
| 编号体系(每级) | 中文数字 | 12 种体系 + 清除编号(多级模式不生效) |
| 分隔符(每级) | 、 |
生成时使用;识别旧编号时兼容 13 种字符分隔符(多级模式不生效) |
| 补全代码块语言 | 关 | 只给缺少语言的围栏代码块补标签 |
| 合并空行 | 关 | 连续 3+ 个换行压成 1 个空行 |
| 块引用首尾补空行 | 关 | 每个 > 引用段首尾各补一行空 >(已有空 > 不重复添加) |
| 插入空白行数 | 20 | 命令「在本段后插入空白行」一次插入的行数(1-100) |
| 自动整理插入图片 | 关 | 笔记保存时把本页新增图片移动到图片文件夹并改写链接(移动而非复制) |
| 图片文件夹 | picture |
相对笔记所在目录,可多级(assets/img);不存在时自动新建 |
| 删除现有分割线 | 关 | 删除 --- / *** / ___,保留 frontmatter 与 setext 下划线 |
| 两个一级标题之间插入 | 关 | 在每个非首个 H1 之前插入 --- |
| 两个二级标题之间插入 | 关 | 每个 H1 分组内、该组非首个 H2 之前插入 ---;无 H1 时按全文处理 |
| 标题语言 | 跟随文档 | 跟随文档(先按正文文字系统判定语言——中/英/日/韩/俄/阿,拉丁字母再按常用词细分西/法/德/葡/意——再要求 AI 用该语言生成;样本太短时交给 AI 判断)/ 指定语言(11 种:中英日韩西法德葡俄意阿) |
| H1-H4 生成内容 | 全开 | emoji / 讨论主题 / 一句话总结,三项至少一项;全关时命令报错并中止 |
| 标题生成提示词 | 中文默认文案 | 对「一句话总结」的生成要求;只读预览 + 修改(弹窗)+ 初始化;清空后按当前界面语言取默认(存的是另一种界面语言的默认文案时同样按当前语言取默认) |
| 主题生成提示词 | 中文默认文案 | 对「讨论主题」的生成要求;只读预览 + 修改(弹窗)+ 初始化;清空后按当前界面语言取默认(存的是另一种界面语言的默认文案时同样按当前语言取默认) |
| AI 服务商 | Claude | 六家:Claude / DeepSeek / Gemini / GPT / Qwen / Kimi;每家内置默认模型 |
| API Key | 空 | 仅用于「重新生成标题」;type=password 掩码显示,可清除;保存在 data.json |
| 体系 | 边界 |
|---|---|
| 中文 / 阿拉伯 / 东阿拉伯 | 阿拉伯与东阿拉伯 1..999 |
| 罗马(大/小写) | 1..3999,超出回退阿拉伯数字 |
| 拉丁(大/小写) | 双射 26 进制(A..Z、AA..AB);解析只接受 1-2 个字母 |
| 圈号 | ①-⑳、㉑-㉟、㊱-㊿;51+ 回退阿拉伯 |
| 谚文 / 带圈谚文 | 各 14 个;超出回退阿拉伯 |
| 西里尔 | 俄语字母表跳 З / О(31 个);超出回退阿拉伯 |
| 片假名 | 伊吕波顺 47 字;超出回退阿拉伯 |