Help & Support

How to use Scrybe

Getting started

Install from the Chrome Web Store

Click Add to Chrome on the store listing. No account needed.

Navigate to any webpage

Open an article, blog post, documentation page, or any content you want to capture.

Click the Scrybe icon

Find Scrybe in your Chrome toolbar. The popup shows the page URL, live stats, and settings.

Choose your output

Click SAVE .MD to download a Markdown file, SAVE .TXT for plain text, or COPY to paste directly into Claude, ChatGPT, or any LLM.

Settings explained

Describe figures

When on, every image is converted to structured text — alt text, inferred type (Chart, Diagram, Logo, etc.), dimensions, and source URL. LLMs can read and reason about these descriptions. When off, images are removed from the output.

Include frontmatter

Adds a metadata header at the top of the document — title, author, published date, source URL, reading time, word count, and estimated token count. Useful for giving LLMs context about the source.

Trim chrome & ads

Removes navigation, sidebars, footers, cookie banners, and ad elements before extraction. Recommended on for cleaner output. Turn off if you notice article content being removed.

Right-click quick capture

Right-click anywhere on a page and select Scrybe this page to instantly download a .md file — no popup needed. Uses Trim chrome & ads by default.

Frequently asked questions

Why does the output show garbled characters?

Open the .md file in VS Code, Obsidian, or Notepad++. These editors detect UTF-8 correctly. Old versions of Windows Notepad may show encoding issues.

Images are showing as 1x1px placeholders.

Some sites use lazy-loading — images only load when scrolled into view. Scroll through the full page before running Scrybe so all images load first.

The extension says "Cannot access this page."

Chrome restricts extensions on certain pages: the Chrome Web Store, chrome:// pages, and some browser-internal pages. This is a Chrome security limitation.

The word count seems too low / content seems missing.

The page may be behind a paywall. Scrybe will show a warning when it detects paywall indicators. Only the publicly accessible portion of the page is extracted.

How is the token count calculated?

Tokens are estimated at approximately 4 characters per token — a standard approximation for English text. It matches well with Claude, GPT-4, and Gemini tokenizers but may vary slightly.

Does Scrybe work on PDFs or local files?

Not currently. Scrybe extracts from live web pages only. PDF and local file support is planned for a future version.

Does Scrybe send my data anywhere?

No. All processing is local — nothing leaves your browser. See the Privacy Policy for full details.

Still need help?

Found a bug or have a feature request? Open an issue on GitHub — it's the fastest way to get a response.

Open an issue on GitHub