cnwenzhihong_Obsidian-auto-paste-link/README.md

Auto Paste Link

Auto Paste Link is an Obsidian plugin that formats a single pasted URL into a Markdown link, image embed, or video embed. For selected supported websites, it can also fill the link title automatically.

中文：这是一个 Obsidian 插件，用于在粘贴单个 URL 时自动生成 Markdown 普通链接、图片嵌入或视频嵌入。对受支持的重要站点，可以自动补全链接标题。

Demo

Open demo video

中文：打开演示视频

Core Behavior

• Normal URL: inserts [](URL) and keeps the cursor inside [].
• Selected text + normal URL: inserts [selected text](URL).
• Supported website URL: inserts [](URL) first, then fills the title when the title request succeeds.
• Image URL: inserts ![](URL).
• Direct video file URL: inserts <video src="URL" controls muted autoplay loop></video>.
• Mixed text is not converted.
• YAML frontmatter is ignored by default.
• Ctrl+Shift+V keeps Obsidian's plain paste behavior and skips plugin processing.

Examples

Normal link:

[](https://www.baidu.com)

Normal link with selected text:

[百度](https://www.baidu.com)

Image link:

![](https://images.steamusercontent.com/ugc/example/hash/?imw=5000)

Video link:

<video src="https://example.com/video.mp4" controls muted autoplay loop></video>

Supported Title Sites

Automatic title fetching is intentionally limited to important supported sites:

bilibili, YouTube, Fab, Steam
TMDb, Letterboxd, Rotten Tomatoes, MyAnimeList, Bangumi, Spotify, Dailymotion
GitHub, GitLab, Gitee, Stack Overflow, Stack Exchange
Reddit, Wikipedia
MDN, npm, PyPI, crates.io, Go Packages
Apple Developer, Unity, Next.js, Vue.js, Node.js
Juejin, CSDN, Tencent Cloud Developer, Alibaba Cloud Developer
WeChat Official Accounts, Douban, CNBlogs, SegmentFault, Jianshu

Other websites can use the generic title fetcher when the title looks clean enough.

Known Unsupported Title Sites

The following sites were tested and did not expose stable, specific titles to normal plugin requests. They may return verification pages, homepage titles, empty titles, or generic app titles. For these sites, copy the title manually or select the title text before pasting the URL.

IMDb, Netflix, AniList
iQIYI, Youku, Mango TV, Maoyan
NetEase Cloud Music, QQ Music
SoundCloud, Twitch, TikTok

Image Detection

Image detection is deliberately conservative. The plugin does not treat every CDN as an image source.

Detection order:

1. Image file extension, such as .jpg, .png, .webp, .gif, .svg, .avif.
2. Trusted image sources.
3. Advanced regular expressions.

Built-in trusted image sources include:

• images.unsplash.com/
• i.imgur.com/
• images.steamusercontent.com/
• pbs.twimg.com/media/
• i.ytimg.com/vi/
• img.youtube.com/vi/
• i0.hdslb.com/bfs/
• i1.hdslb.com/bfs/
• i2.hdslb.com/bfs/

Generic CDN domains such as cloudfront.net, akamaihd.net, fastly.net, and googleusercontent.com are not built in, because they can serve images, webpages, scripts, downloads, and many other resource types.

Custom Trusted Image Sources

Use custom trusted image sources for hosts or paths that clearly serve images.

Each source has:

• Host: required, for example images.example.com.
• Path prefix: optional, for example /media/.
• Include subdomains: optional. Keep it off unless subdomains are also known image sources.

Examples:

host: images.example.com
path prefix: /media/
include subdomains: off

This matches:

https://images.example.com/media/abc

It does not match:

https://images.example.com/files/abc
https://sub.images.example.com/media/abc

If include subdomains is enabled, then subdomains such as cdn.images.example.com can also match.

Advanced Regular Expressions

Advanced image URL regular expressions are still available for edge cases, but they should be treated as an escape hatch.

Use trusted image sources first. Use regular expressions only when a source cannot be expressed by host and path prefix.

The default advanced rule recognizes explicit image format query parameters, for example:

https://example.com/resource?id=1&format=jpg

Settings

• Auto fetch supported site titles: enabled by default.
• Title request timeout: 3000ms by default. Fab is usually slower than other supported sites.
• Use selected text as the link title: enabled by default.
• Process YAML frontmatter: disabled by default.
• Auto embed image links: enabled by default.
• Add newline after image links: enabled by default.
• Image extensions are configurable.
• Custom trusted image sources are configurable.
• Advanced image URL regular expressions are configurable.
• Auto embed video links: enabled by default.
• Video extensions are configurable. The default is mp4.

Manual Installation

1. Download main.js, manifest.json, and styles.css from the latest GitHub release.
2. Create this folder in your vault: .obsidian/plugins/auto-paste-link/.
3. Put the three downloaded files into that folder.
4. Enable the plugin in Obsidian settings.

Development

corepack pnpm install
corepack pnpm test
corepack pnpm build

The GitHub release tag must exactly match the version in manifest.json, for example 1.0.4 without a v prefix. Release assets must include main.js, manifest.json, and styles.css.

Adding Supported Sites With AI

Only add a site-specific provider when the generic title fetcher returns a bad title and the site has a stable way to produce a cleaner one.

1. Pick the closest group file in src/core/titleProviders/groups/.
2. Add a small TitleProvider with id, displayName, matches, createRequests, and parse.
3. Prefer stable APIs or oEmbed endpoints when available. Otherwise clean a predictable suffix from og:title, twitter:title, or <title>.
4. Do not add sites that only return verification pages, login pages, homepage titles, or vague app titles.
5. Register a new group in src/core/titleProviders/providers.ts only when no existing group fits.
6. Add tests in tests/core.test.ts for host matching and title cleanup.
7. Update the supported or unsupported site list in this README.
8. Run:

node --experimental-default-type=module --test tests/*.test.ts
./node_modules/.bin/tsc.CMD -noEmit -skipLibCheck
node esbuild.config.mjs production