How-to · Productivity · By Productivity and Collaboration

How to use Kreuzberg Document Extractor in your AI agent

Install and use the Kreuzberg Document Extractor skill in Claude Code, Cursor, Codex, Gemini CLI, OpenCode, and Windsurf — with sample prompts.

What you'll get

Kreuzberg extracts content from nearly any document format you throw at it—PDFs, Word files, Excel sheets, PowerPoint decks, images (with OCR), emails, HTML, and archives. It's built for AI agents to read files on your behalf so you can ask 'What's in this contract?' or 'Summarize this scanned receipt' without manually opening anything. Perfect for founders, PMs, and operators who need AI to digest documents quickly.

Step 1 — Get the skill source URL

The Kreuzberg Document Extractor skill lives at:

Open ↗

Step 2 — Add it to your AI agent

Copy one of these prompts and paste it into your AI tool. The agent will fetch and install the skill for you.

Claude Code

Add the agent skill from https://github.com/kreuzberg-dev/kreuzberg/tree/main/skills/kreuzberg to my project.

Cursor

Install this agent skill: https://github.com/kreuzberg-dev/kreuzberg/tree/main/skills/kreuzberg

Codex

Add this skill to .codex/skills/: https://github.com/kreuzberg-dev/kreuzberg/tree/main/skills/kreuzberg

Gemini CLI

Add the skill at https://github.com/kreuzberg-dev/kreuzberg/tree/main/skills/kreuzberg to my .gemini/skills/ directory.

OpenCode

Install this agent skill from https://github.com/kreuzberg-dev/kreuzberg/tree/main/skills/kreuzberg

Windsurf

Add this skill: https://github.com/kreuzberg-dev/kreuzberg/tree/main/skills/kreuzberg

Step 3 — Try these prompts

Once the skill is installed, ask your AI things like:

  • You need AI to read contracts, invoices, or research papers and pull out key info
  • Processing receipts, scanned documents, or photos with text and need OCR
  • Bulk-extract data from a folder of mixed file types (PDFs, Word docs, spreadsheets)
  • Building a knowledge base and need to ingest documents automatically

What this skill does

  • Extracts text, tables, metadata, and images from 90+ file formats including PDF, Word, Excel, PowerPoint, images, emails, and HTML
  • Runs OCR on scanned PDFs and photos so AI can read text from images or receipts
  • Processes batches of files at once—upload a folder of contracts and get all the data out
  • Outputs clean Markdown or JSON for easy reading and further AI processing
  • Handles password-protected PDFs and multi-language documents
SkillKreuzberg Document Extractor
Made byProductivity and Collaboration
CategoryProductivity
SourceView original ↗