Claude Code Plugins

Community-maintained marketplace

Feedback

fetch-text

@bdambrosio/Cognitive_workbench
8
0

Fetch all text from URL or base64 PDF. Collection-aware (extracts first item if given Collection). Auto-detects format (PDF/HTML/MD/TXT) and extracts complete text content.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name fetch-text
description Fetch all text from URL or base64 PDF. Collection-aware (extracts first item if given Collection). Auto-detects format (PDF/HTML/MD/TXT) and extracts complete text content.
type python
schema_hint [object Object]
parameters [object Object]
examples {"type":"fetch-text","target":"https://example.com/doc.pdf","out":"$text"}, {"type":"fetch-text","target":"https://example.com/page.html","out":"$text"}, {"type":"project","target":"$papers","fields":["metadata.uri"],"out":"$urls"}, {"type":"fetch-text","target":"$urls","out":"$text"}

Fetch Text

Fetches complete text content from URLs or base64 PDFs. Auto-detects format (PDF/HTML/Markdown/Text) and extracts all text without filtering.

Collection-aware: If given a Collection ID, extracts the first Note's content field and uses it as the URL. IMPORTANT: The Collection must contain Notes where content is a URL string. For structured Notes (e.g., from semantic-scholar or search-web), use project first to extract URLs into a Collection of URL strings.

Input

  • url_or_content: URL string, base64-encoded PDF content, or Collection ID (uses first item's content field as URL)

Output

Returns structured JSON with:

  • text: Full extracted text content (all text, no filtering)
  • format: Detected format ("pdf", "html", "markdown", "text")
  • metadata: Format-specific metadata (source_url, pdf_metadata for PDFs, html_metadata for HTML)
  • page_count: Number of pages (PDF only)
  • char_count: Total character count

Usage

Direct URL fetch:

{"type":"fetch-text","target":"https://arxiv.org/pdf/1706.03762.pdf","out":"$paper_text"}

HTML page:

{"type":"fetch-text","target":"https://example.com/article.html","out":"$article_text"}

Collection input (extract URLs first):

{"type":"project","target":"$papers","fields":["metadata.uri"],"out":"$urls"}
{"type":"fetch-text","target":"$urls","out":"$paper_text"}

Differences from search-web

  • fetch-text: Gets ALL text from a single URL (no filtering, no search)
  • search-web: Searches web, returns FILTERED excerpts from multiple URLs (query-relevant snippets)

Use fetch-text when you have a specific URL and want complete content. Use search-web when searching for information.