Open the app — HWP Image Detector
Sometimes you have to check every photograph across dozens of reports.
Clearing copyright, building a list of plates, auditing who inserted which figure.
Your only option is opening each file and scrolling through it.
And when you find a picture you still have to read around it to remember what it was.
This tool takes a whole folder and sweeps every HWP/HWPX file inside.
For each document it shows the embedded images along with the sentences before and after where that image sits.
It puts the context a picture alone can’t give you right next to it.
Hangul doesn’t need to be installed, and no file leaves your machine.
How to Use
- Set Context characters (how many characters to show before/after the image; default 15).
- Click Select folder & analyze and choose the target folder. Subfolders are included recursively.
- When the progress bar finishes, the result appears. Use Download HTML to save it, or Print to print it.
For security reasons the browser only grants folder access when you pick the folder yourself; paths are never read automatically.
Supported formats
| Format | Supported | Notes |
|---|---|---|
HWP 5.x (.hwp) | Yes | Parses the OLE compound-file structure directly |
HWPX (.hwpx) | Yes | Parses the ZIP+XML (OWPML) structure directly |
| HWP 3.x or older · password-protected | No | Not parsed; listed under Errors in the result |
The result
- Summary: number of files with images, total images, and files with errors.
- Images per file: a table of file name (click to jump to the image), count, and path.
- Errors: files that could not be parsed and why. One bad file never stops the whole run.
- Image cards: preceding context on the left, the actual image in the middle, following context on the right.
Surrounding context
This is the heart of the tool. A pile of pictures on its own tells you nothing about what they are.
The text is cut from the document flow before and after the point where the image is anchored. Images often sit in their own empty paragraph, so the context comes from the neighboring paragraphs, not the same one. When one side has no text (e.g. an image at the very start or end), it shows (no preceding text) / (no following text). Embedded images that are not placed in the body are shown separately without context.
Variant Hanja in the context are normalized to standard forms, and special characters that render as tofu (□) are removed.
Download / Print
- Download HTML: saves the same result — images included — as a single self-contained HTML file (
hwp-image-report.html) that opens on its own with no external dependencies. - Print: opens the browser print dialog and prints only the result (title, settings, buttons and background are excluded).
Good to Know
Notes
- Latest Chrome / Edge is recommended.
- Very large folders (thousands to tens of thousands of files with many images) can use a lot of browser memory and make the result heavy.
- The count reflects images that actually appear in the document (inserting the same image twice counts as two).
Internationalization
The whole page supports Korean / English; toggle KO/EN in the top-left to switch instantly.
Open the app — HWP Image Detector