url_extract_images - Extract and Host Images via GitHub

Author: dwdecon
GitHub Repository: https://github.com/dwdecon/url_extract_images
Dify Marketplace: (optional, if published)

One-stop Dify tool to render pages (including JS-heavy sites), extract images from HTML/PDF URLs, filter by size, and upload them to your GitHub repository so you get stable, shareable URLs.

Description

Tool	Input	Output
url_extract_images	Web page or PDF URL	Hosted GitHub URLs plus detected titles/alts

GitHub Setup (Step-by-Step)

1. Create repository

   • GitHub -> New -> name it (e.g., image-bed).
   • Visibility: Public if you want hosted URLs accessible to anyone; Private if you need restricted access.
   • Default branch: typically main.

2. Create Personal Access Token (classic)

   • Settings -> Developer settings -> Personal access tokens -> Tokens (classic) -> Generate new token.
   • Scopes: enable repo (covers contents read/write).
   • Copy the token (looks like ghp_xxx...) and keep it secret.

3. Fine-grained token (alternative)

   • Developer settings -> Fine-grained tokens -> Generate new token.
   • Select the target repository; grant Contents read/write.
   • If your org enforces SSO, click Authorize for that org.

4. Store credentials in Dify

   • github_token: the PAT above.
   • github_owner: your GitHub username or org.
   • github_repo: repository name.
   • github_branch: branch to write to (e.g., main).
   • github_path (optional): folder prefix like images/.

5. Validate access

   • Ensure the repo exists and the PAT has write access.
   • 404/403 usually means wrong owner/repo, missing repo scope, private repo without access, or SSO not authorized.

Installation

• Upload the packaged release (e.g., url_extract_images-0.3.0.difypkg) from the Dify Plugins page.
• For remote install, set environment variables: INSTALL_METHOD=remote, REMOTE_INSTALL_URL, REMOTE_INSTALL_KEY.

Configuration

Provider credentials (stored in Dify):

• github_token: PAT with repo scope.
• github_owner: repo owner (user or org).
• github_repo: repository name.
• github_branch: branch to write to (default main).
• github_path: optional folder prefix, e.g., images/.

Tool parameters (per invocation):

• url (required): page or PDF URL to extract from.
• max_images (default 20): max images to return.
• min_width / min_height (default 100 px): size filter.
• timeout (default 30 s): page load timeout.

Usage (Dify Workflow)

1. Add a Tool node and pick url_extract_images.
2. Bind the url input (direct URL or a workflow variable).
3. Optionally adjust max_images, size filters, and timeout.
4. Run the workflow—the tool uploads images to GitHub and returns cloud URLs plus detected titles/captions.

Changelog

• 0.3.0 - Renamed tool to url_extract_images, refreshed docs, kept GitHub upload flow and image extraction logic.
• 0.2.x - Initial releases of url_to_images with GitHub upload support and size filtering.

Privacy

• GitHub PAT is stored as a secret credential in Dify; it is used only to access your target repo/branch/path.
• URLs you provide (and extracted images) are uploaded to your GitHub repository; choose repo visibility to match sensitivity.
• The tool fetches the target URL for rendering/downloading; avoid sending pages with data you do not want fetched.
• No telemetry or third-party analytics beyond the GitHub API calls needed for uploads.