How to Set Up Apache Tika with Open WebUI in 2026

If you've tried using OpenWebUI's default document parsing — like uploading a PDF and asking questions — you probably noticed it's not great out of the box. Let's fix that with Apache Tika.

The problem? OpenWebUI by default doesn't use specialized tools for extracting text from complex documents. So if your input is a scanned PDF, a DOC file, or even just a weirdly formatted PDF, it will likely fail or produce garbage output.

Quick Answer

Run Apache Tika as a separate container next to Open WebUI, then point Open WebUI's document extraction setting at http://tika:9998/tika. If you host Open WebUI on a Docker platform like Sliplane, deploy Tika as another service in the same project and use the internal service URL instead of exposing Tika publicly.

Here is a 1 minute tutorial on how to set up Apache Tika with OpenWebUI:

Why Use Apache Tika?

Apache Tika is a mature open-source project by the Apache Foundation. It's been around for over 18 years and handles hundreds of document formats, including:

• PDF
• DOC/DOCX
• PPT/PPTX
• XLS/XLSX
• EPUB
• HTML
• Plaintext
• Legacy formats like DOC (97-2003)
• Many more

Tika extracts text and metadata reliably, making it a much better backend for document parsing in OpenWebUI.

How to Set Up Apache Tika with OpenWebUI

Docker Compose Setup

Here's a simple compose.yml file that runs both OpenWebUI and Apache Tika:

services:
  openwebui:
    image: ghcr.io/open-webui/open-webui:main
    ports:
      - "3000:8080"
    volumes:
      - openwebui:/app/backend/data

  tika:
    image: apache/tika:latest
    ports:
      - "9998:9998"

volumes:
  openwebui:

Configure OpenWebUI

1. Start both services with docker compose up -d
2. Open OpenWebUI at http://localhost:3000
3. Go to Admin Panel → Settings → Documents
4. Set the Content Extraction Engine to Tika and then the URL to:

http://tika:9998/tika

5. Save the settings
6. Validate the setup by uploading a document and checking the Tika logs.

You should see something like:

INFO  [qtp1393198164-60] 15:03:46,042 org.apache.tika.server.core.resource.TikaResource /tika (application/pdf)

Example: Parsing a PDF

Upload a PDF in OpenWebUI, ask questions about its content, and compare the results before and after setting up Tika. You'll notice:

• Better text extraction
• Support for more formats
• Fewer parsing errors

Try uploading complex documents like:

• Scanned PDFs
• Multi-column layouts
• Documents with tables
• Legacy Microsoft Office files

Troubleshooting

If documents aren't parsing correctly:

1. Check Tika is running:
   docker ps | grep tika
   

2. Test Tika directly:
   curl -X PUT --data-binary @your-document.pdf http://localhost:9998/tika --header "Content-type: application/pdf"
   

3. Check OpenWebUI logs:
   docker logs <openwebui-container-id>
   

FAQ

Is there an alternative to Apache Tika? Yes, but most alternatives are either commercial or more limited. Tika is free, open source, and extremely battle-tested.

Does this cost anything? No. Both OpenWebUI and Apache Tika are open source and free to use.

Does it scale? For most use cases, a single Tika container is fine. If you need to scale, you can run multiple Tika containers behind a load balancer.

What about security? Tika processes documents server-side, so ensure your setup is properly secured if handling sensitive documents. Consider running it in an isolated environment.

Video Demo

Here's a 40-second demo video showing:

• How to deploy Apache Tika on Sliplane
• How to configure it in OpenWebUI
• How to test it with a real PDF

Conclusion

Setting up Apache Tika with OpenWebUI transforms document parsing from frustrating to functional. With just a few configuration steps, you can handle hundreds of document formats reliably.

If you have any questions, feel free to reach out in our Support chat (bottom right).

Cheers, Jonas