Open WebUI¶

Kreuzberg can serve as the Content Extraction Engine for Open WebUI, replacing docling-serve or other extraction backends. When users upload documents in Open WebUI, Kreuzberg extracts the text — supporting 91+ document formats and 248 programming languages — including PDFs, Office documents, images with OCR, and source code — without requiring a GPU.

How it works¶

flowchart LR
    Upload[User uploads document] --> OpenWebUI[Open WebUI]
    OpenWebUI --> Kreuzberg[Kreuzberg API Server]
    Kreuzberg --> Extract[Extract text + OCR]
    Extract --> Markdown[Markdown output]
    Markdown --> OpenWebUI
    OpenWebUI --> RAG[Vector store / RAG]

    style Kreuzberg fill:#87CEEB
    style OpenWebUI fill:#FFD700
    style RAG fill:#90EE90

1. A user uploads a document in Open WebUI.
2. Open WebUI sends the file to Kreuzberg's API server.
3. Kreuzberg extracts the content (with OCR for images/scanned PDFs) and returns markdown.
4. Open WebUI stores the text in its vector database for retrieval-augmented generation.

Supported engine modes¶

Kreuzberg implements two Open WebUI-compatible endpoints. Choose either mode:

Docling engine (recommended)¶

Kreuzberg implements the docling-serve API at POST /v1/convert/file, acting as a drop-in replacement.

External engine¶

Kreuzberg also implements the generic external document loader API at PUT /process.

Both modes can also be configured via the Open WebUI Admin UI at Settings > Documents > Content Extraction Engine.

Quick start with Docker Compose¶

services:
  kreuzberg:
    image: ghcr.io/kreuzberg-dev/kreuzberg:latest-core
    ports:
      - "8000:8000"
    command: ["serve", "--host", "0.0.0.0", "--port", "8000"]
    healthcheck:
      test: ["CMD", "kreuzberg", "version"]
      interval: 10s
      timeout: 5s
      retries: 5

  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    ports:
      - "3000:8080"
    environment:
      CONTENT_EXTRACTION_ENGINE: "docling"
      DOCLING_SERVER_URL: "http://kreuzberg:8000"
    depends_on:
      kreuzberg:
        condition: service_healthy

docker compose up

Open http://localhost:3000 and upload a document. Kreuzberg handles the extraction.

Running Kreuzberg standalone¶

If Open WebUI is already running, you only need Kreuzberg:

docker run -p 8000:8000 ghcr.io/kreuzberg-dev/kreuzberg:latest-core \
  serve --host 0.0.0.0 --port 8000

kreuzberg serve --host 0.0.0.0 --port 8000

Then point Open WebUI to http://<kreuzberg-host>:8000 using either engine mode.

Verifying the integration¶

Test the endpoints directly:

curl -F "files=@document.pdf" http://localhost:8000/v1/convert/file

{
  "document": {
    "md_content": "# Extracted content..."
  },
  "status": "success"
}

curl -X PUT \
  -H "Content-Type: application/pdf" \
  -H "X-Filename: document.pdf" \
  --data-binary @document.pdf \
  http://localhost:8000/process

{
  "page_content": "# Extracted content...",
  "metadata": {
    "source": "document.pdf"
  }
}

Supported formats¶

Kreuzberg supports 91+ file formats and 248 programming languages — including PDF, DOCX, PPTX, XLSX, images (PNG, JPG, TIFF with OCR), HTML, Markdown, email (EML, MSG), archives (ZIP, TAR), ebooks (EPUB), source code, and more. See the full format list.