外掛標籤
開發者團隊
② 後台搜尋「Imgsoft Localize External Images」→ 直接安裝（推薦）
原文外掛簡介
Images hotlinked from other websites slow your pages down, break when the other site removes them, and can disappear from your posts without notice. Imgsoft Localize External Images finds them, shows you exactly what it found, and moves the ones you choose into your own Media Library.
It can also turn text links to images — for example the ones ChatGPT puts into answers — into the images themselves.
Nothing changes until you click “Localize”. Before every change the plugin saves a backup of the post, so any localization can be rolled back.
Scan
Scans all posts and pages, including drafts, scheduled and private ones.
Finds images in , srcset,
Shows every image with a preview, its alt text and the post it belongs to. Everything is selected by default; untick what you want to keep external.
Search, filters by post type and status, pagination — works on sites with tens of thousands of posts.
Localize
Downloads each image once; if the same image is used in several posts, the file is reused.
Detects the real file type by content, so links without an extension (for example image CDNs) work too.
Recognizes well-known image hosts (Unsplash, Imgur, ChatGPT, Google previews and others) even when an address has no file extension — in styles, block data and links too.
Size limit for downloaded files (5 MB by default, adjustable or off): larger files are not downloaded and go to problem links, where you can load a single file ignoring the limit.
Saves each post once, no matter how many images it contains — one revision instead of one per image.
SVG can be localized safely: each file is cleaned from scripts, event handlers and external references before it is saved (optional, off by default). Manual SVG uploads stay forbidden.
History and rollback
Every localization is a session with its own backups.
Roll back only the links the plugin changed (your later manual edits stay), or restore the whole post to its state before the plugin ran.
Roll back a single link, a post, or a whole session. Optionally delete the downloaded files if nothing else uses them.
See how much space the backups take and delete archives you no longer need.
Problem links
Links that cannot be downloaded are moved to a separate list and sorted by reason:
Broken — 404, 410, the domain does not exist, or there is no image at the address.
Temporarily unavailable — timeout, server error, too many requests.
Download forbidden — the site refuses to give the file to a server (403, 401).
File type not accepted — the image downloads but WordPress does not accept its type.
Too large — the file is bigger than the size limit set in the settings.
After each scan the plugin quickly rechecks them and brings back the ones that became available. From the list you can run a longer recheck, replace a link with a new address (it is checked before anything changes), pick an image from your Media Library, or exclude the link. The alt text is shown with a one-click “find similar” search.
Converting links into images
Text links to images, like Photo of the beach, can be turned into the images themselves. The image is downloaded to your Media Library first, and the link text becomes its alt.
Two ways to use it: as an option of the main scan, or with the separate “Links → images” tool that handles only links and does not touch ordinary external images.
Ready-made sets that work right after installation: ChatGPT images, Google image previews, photo stocks and image hosting, social networks, Wikimedia Commons, or all of them together.
Wikimedia Commons is handled with its licenses: public domain files are inserted as is, CC BY / CC BY-SA 4.0 files link to their source page, other licenses also get an author and license caption.
Your own rules: conditions joined with “and” (exactly matches, contains, starts with, does not contain, domain, regular expression, inside tag, not inside tag), exclusions, excluded posts, and a “check an address” tool.
Links inside image captions and links to an image that is already shown in the post are never turned into a second image.
If a link cannot be downloaded because it is broken or the server is down, it stays a link. If the server only forbids downloading, the image is shown from the external address.
Rule sets and configuration
Save, rename and delete your own sets of rules, with comments; apply a set or add it to the current rules.
Export and import sets as files — as a full snapshot or as rules only, without the list of posts.
Configuration editor: the whole configuration as text with a built-in description of the format. You can send it to an AI assistant with a description of what you need, paste the answer back, check it and apply. Errors are shown with their exact place, and you see what will change before applying.
The previous settings are saved automatically before every change, so you can always go back.
Exclusions
Exclude a single image, a whole domain (with subdomains) or an entire post or page.
Exclude straight from the scan list, from a dedicated “scan and add to exclusions” view, or add addresses manually.
Regular expressions (“regex: /avatars/” or “/(gravatar|avatars)/i”) and tags (“tag: figcaption”, ”
