Kreuzberg MCP — Anwendungsfälle, Installieren & Live-Demo

Warum nutzen

Hauptfunktionen

97+ Formate — PDF, DOCX, XLSX, PPTX, Bilder, HTML, EPUB, RTF
Rostkern — schneller, geringer Speicherbedarf im Vergleich zu Python-Alternativen
Eingebaute OCR (Tesseract/PaddleOCR) für gescannte Dokumente
Preserves structure — tables as markdown, headings, lists
Extracts metadata: author, creation date, word count, language

Live-Demo

In der Praxis

kreuzberg.replay ▶ bereit

0/0

Installieren

Wählen Sie Ihren Client

~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "kreuzberg": {
      "command": "uvx",
      "args": [
        "kreuzberg-mcp"
      ]
    }
  }
}

Öffne Claude Desktop → Settings → Developer → Edit Config. Nach dem Speichern neu starten.

~/.cursor/mcp.json · .cursor/mcp.json

{
  "mcpServers": {
    "kreuzberg": {
      "command": "uvx",
      "args": [
        "kreuzberg-mcp"
      ]
    }
  }
}

Cursor nutzt das gleiche mcpServers-Schema wie Claude Desktop. Projektkonfiguration schlägt die globale.

VS Code → Cline → MCP Servers → Edit

{
  "mcpServers": {
    "kreuzberg": {
      "command": "uvx",
      "args": [
        "kreuzberg-mcp"
      ]
    }
  }
}

Klicken Sie auf das MCP-Servers-Symbol in der Cline-Seitenleiste, dann "Edit Configuration".

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "kreuzberg": {
      "command": "uvx",
      "args": [
        "kreuzberg-mcp"
      ]
    }
  }
}

Gleiche Struktur wie Claude Desktop. Windsurf neu starten zum Übernehmen.

~/.continue/config.json

{
  "mcpServers": [
    {
      "name": "kreuzberg",
      "command": "uvx",
      "args": [
        "kreuzberg-mcp"
      ]
    }
  ]
}

Continue nutzt ein Array von Serverobjekten statt einer Map.

~/.config/zed/settings.json

{
  "context_servers": {
    "kreuzberg": {
      "command": {
        "path": "uvx",
        "args": [
          "kreuzberg-mcp"
        ]
      }
    }
  }
}

In context_servers hinzufügen. Zed lädt beim Speichern neu.

claude mcp add kreuzberg -- uvx kreuzberg-mcp

Einzeiler. Prüfen mit claude mcp list. Entfernen mit claude mcp remove.

Anwendungsfälle

Praxisnahe Nutzung: Kreuzberg

Extract tables from a messy PDF into clean markdown

👤 Analysts dealing with report PDFs ⏱ ~10 min beginner

Wann einsetzen: You have a PDF with tables that pdftotext mangles and you don't want to retype them.

Voraussetzungen

MCP installed — uvx kreuzberg-mcp — or add via claude mcp add

Ablauf

Extract

Use kreuzberg to extract /docs/2025-annual-report.pdf. Give me the tables as markdown and the body text separately.✓ Kopiert

→ Clean markdown tables with preserved headers
Verify

For the "Revenue by Segment" table, reconcile the column totals. Flag any OCR misreads.✓ Kopiert

→ Arithmetic check with flagged cells

Ergebnis: Markdown tables you can paste into a doc without rework.

Fallstricke

Scanned PDF — OCR mistakes 6 for 8 — Use the OCR confidence output and re-scan low-confidence cells manually

Kombinieren mit: filesystem

Ingest a folder of mixed-format docs for downstream indexing

👤 Engineers building RAG pipelines ⏱ ~30 min intermediate

Wann einsetzen: Client drops you a zip with PDFs, Word docs, PowerPoints, and you need clean text for embedding.

Voraussetzungen

Filesystem MCP scoped to the folder — Launch fs MCP with the ingest dir as root

Ablauf

Inventory

List every file under /ingest/. For each, call kreuzberg.detect_format and report.✓ Kopiert

→ Format-per-file table
Bulk extract

For each file, extract text + metadata. Write cleaned .txt next to the original and a manifest.json with metadata.✓ Kopiert

→ All files processed; manifest contains every entry
Quality check

List every file where extraction returned <100 chars — those are likely scanned or corrupt. Re-run with OCR forced.✓ Kopiert

→ Low-content files identified and retried

Ergebnis: A folder of cleaned text files ready for embedding, with a metadata manifest.

Fallstricke

Encrypted PDF — Kreuzberg returns an error — decrypt with qpdf or ask for the unlocked copy

Kombinieren mit: filesystem · memory

Kombinationen

Mit anderen MCPs für 10-fache Wirkung

kreuzberg + filesystem

Walk a folder and extract every doc in place

For each PDF under /docs, extract text and save as .md next to it.✓ Kopiert

kreuzberg + memory

Ingest extracted content into a knowledge graph

Extract /contracts/*.pdf and store key terms in memory for cross-doc querying.✓ Kopiert

Werkzeuge

Was dieses MCP bereitstellt

Werkzeug	Eingaben	Wann aufrufen	Kosten
extract_text	path: str, ocr?: bool	Primary extraction call	free
extract_metadata	path: str	Just want metadata without body text	free
extract_tables	path: str	Table-focused extraction	free
detect_format	path: str	Confirm format before extraction	free

Kosten & Limits

Was der Betrieb kostet

API-Kontingent: Unlimited — local
Tokens pro Aufruf: Proportional to doc size; a 20-page PDF ≈ 8k tokens of output
Kosten in €: Free (open source)
Tipp: Use extract_metadata first on large files to avoid processing irrelevant ones

Sicherheit

Rechte, Secrets, Reichweite

Credential-Speicherung: None for local mode

Datenabfluss: None — all processing is local

Malformed PDFs can trigger parser edge cases — sandbox the MCP if processing untrusted uploads

Fehlerbehebung

Häufige Fehler und Lösungen

ModuleNotFoundError: tesseract

Install Tesseract system binary: brew install tesseract / apt install tesseract-ocr

Prüfen: `tesseract --version`

Empty output on PDF

Likely image-only PDF — re-run with ocr=true

Prüfen: Check output.metadata.has_text_layer

XLSX tables come out jumbled

Pass sheet name explicitly: tool supports sheet arg

Alternativen

Kreuzberg vs. andere

Alternative	Wann stattdessen	Kompromiss
markdownify-mcp	You want a lighter, Node-based converter without OCR	Fewer formats, no table preservation
Unstructured.io	You need enterprise PDF parsing and accept the cloud cost	Paid; cloud-hosted

Mehr

Ressourcen

📖 Offizielle README auf GitHub lesen

🐙 Offene Issues ansehen

🔍 Alle 400+ MCP-Server und Skills durchsuchen