How to Deploy RagApp with Docker — A Step-by-Step Tutorial

📋 How to Deploy RagApp with Docker in Under 5 Minutes

I'm going to walk you through exactly how I set up RagApp — from zero to chatting with my documents. No prior RAG experience needed. If you have Docker installed, you're 90% of the way there.

Step 1: Pull and Run (30 seconds)

This is embarrassingly simple. Open a terminal and run:

docker run -p 8000:8000 ragapp/ragapp:latest

That's it. While it downloads (the image is about 1.2GB — grab a coffee if you're on slow internet), let me explain what's happening. The container starts a Python backend (LlamaIndex) and serves both the Admin UI and Chat UI on port 8000.

Wait until you see something like Uvicorn running on http://0.0.0.0:8000 in the logs. That's your signal it's ready.

Step 2: Configure in Admin UI (2 minutes)

Open http://localhost:8000/admin in your browser. You'll see the configuration dashboard.

Choose Your LLM

You need at least one LLM provider configured. Here's what I tested:

Pro tip: I spent 10 minutes debugging why Ollama wasn't working. The issue? RagApp in Docker needs to reach Ollama's host. On Linux, use --add-host host.docker.internal:host-gateway or set the Ollama URL to http://172.17.0.1:11434. Don't make my mistake!

Add Your Data

Point RagApp at your documents. It supports PDFs, text files, markdown, and web URLs. I threw a mix of technical whitepapers and internal documentation at it — the chunking handled it well out of the box. If you want to tweak chunk size or overlap, the Admin UI has sliders for that.

Step 3: Chat with Your Data (1 minute)

Once configured, navigate to http://localhost:8000 (no /admin). You'll see the Chat UI.

Ask a question about your documents. Something like:

"What were the key findings from the Q3 technical review?"

If you see a coherent, sourced answer — congratulations, your RAG is working. If not, double-check that you actually ingested documents (this got me — I configured everything but forgot to upload files).

Step 4 (Optional): Deploy with Docker Compose

For a more production-like setup with volume persistence:

services:
  ragapp:
    image: ragapp/ragapp:latest
    restart: unless-stopped
    ports:
      - 8080:8080
    volumes:
      - ./data/ragapp:/data

I use port 8080 here intentionally — avoids conflicts if you already have something on 8000. The /data volume persists your configuration and document index across container restarts.

Common Pitfalls I Hit

• "Connection refused" on Ollama — See the pro tip above. Docker networking is the issue, not RagApp.
• No answers after configuration — You need to actually upload documents in the Admin UI, not just set up the LLM. Embarrassingly, I missed this step.
• Slow first query — The first question after startup takes ~5-10 seconds as the embedding model initializes. Subsequent queries are near-instant.

Final Thoughts

RagApp isn't trying to be the most advanced RAG platform out there. What it does well is remove friction. Within 5 minutes of deciding to try it, I was getting answers from my documents. For internal tools, demos, and small-team knowledge bases, that's exactly what you need.