How to Run Ollama on a QNAP NAS - Setup Guide 2026

Step 1: Install Container Station

Container Station is QNAP's Docker runtime, available through the App Center.

1. Log into the QNAP QTS web interface.
2. Open the App Center.
3. Search for Container Station.
4. Click Install. The application is free.
5. Once installed, open Container Station from the QTS main menu.

Container Station on QTS uses a slightly different interface from Docker Desktop, but the underlying engine is standard Docker CE. Containers, images, and Docker Compose files work identically to any other Docker host.

Step 2: Create Shared Folders for Container Data

Create dedicated folders for Ollama model storage and Open WebUI data before deploying containers.

1. Open File Station in QTS.
2. Navigate to your main data volume (typically DataVol1).
3. Create a folder named docker.
4. Inside docker, create two subfolders: ollama and open-webui.

If your QNAP has an NVMe M.2 SSD installed and configured as a cache or storage volume, storing the ollama folder there significantly reduces model load times. NVMe storage can reduce model-loading time, but the result varies with model size, cache state, storage and RAID configuration, concurrent I/O, and available RAM. The NVMe does not affect inference speed once the model is in RAM.

Step 3: Deploy Ollama and Open WebUI via Docker Compose

Container Station supports Docker Compose projects. This is the cleanest deployment method.

1. Open Container Station.
2. Click Applications in the left sidebar (the Docker Compose interface).
3. Click Create.
4. Name the application ollama.
5. Paste the following Docker Compose configuration:

version: '3.8'
services:
  ollama:
    image: ollama/ollama:latest
    container_name: ollama
    restart: unless-stopped
    ports:
      - "11434:11434"
    volumes:
      - /share/docker/ollama:/root/.ollama
    environment:
      - OLLAMA_HOST=0.0.0.0
      - OLLAMA_NUM_PARALLEL=1
      - OLLAMA_MAX_LOADED_MODELS=1

  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    restart: unless-stopped
    ports:
      - "3000:8080"
    volumes:
      - /share/docker/open-webui:/app/backend/data
    environment:
      - OLLAMA_BASE_URL=http://ollama:11434
    depends_on:
      - ollama

QNAP documents shared-folder bind mounts as /share/{SHARED_FOLDER_NAME}; if docker is the shared folder, use /share/docker/ollama and /share/docker/open-webui. Adjust the path if your volume is named differently. OLLAMA_MAX_LOADED_MODELS=1 ensures only one model is loaded at a time, which is important on NAS hardware with limited RAM.

Click Validate and then Create. Container Station pulls the images and starts both containers. Image pull takes 3-10 minutes depending on connection speed.

Step 4: Pull AI Models

Use Container Station's terminal to pull models into the Ollama container.

1. In Container Station, click Containers.
2. Find the ollama container and click it.
3. Click the Terminal icon (command prompt).
4. In the terminal, run: ollama pull llama3.2:3b

Alternatively, use SSH:

1. SSH into the QNAP: ssh admin@your-qnap-ip
2. Run: docker exec -it ollama ollama pull llama3.2:3b

Recommended models for QNAP hardware (by use case):

• llama3.2:3b. Fast, low RAM, good general assistant. Start here.
• mistral:7b. Strong for structured tasks, coding, and analysis. Same RAM requirement as Llama 7B.
• qwen2.5-coder:7b. Best for code review and generation at the 7B tier.
• phi3:mini. 3.8B, very efficient for its size, fast response on QNAP hardware.

Run ollama list in the container terminal to see all installed models and their sizes.

Step 5: Access Open WebUI