How to Set Up Ollama AI in Open Web UI with Docker Desktop

-
Ollama and Docker Setup Hero

Local AI Revolution: Deploying Ollama & Open Web UI with Docker

Running Large Language Models (LLMs) locally offers privacy, zero latency, and freedom from subscription fees. In this guide, we won't just install Ollama; we will build a full local stack using Docker to connect a powerful backend with the sleek Open Web UI frontend.

⚙️ The Architecture: How It Works

Before pasting commands, it is crucial to understand the setup. We are creating two distinct Docker containers that need to communicate:

  • Container A (Backend): Runs Ollama, serving the API on port 11434.
  • Container B (Frontend): Runs Open Web UI, accessible via your browser on port 3000.
  • The Bridge: We use the host.docker.internal flag to allow Container B to "see" Container A safely.

📋 Prerequisites

Ensure your environment is ready to handle LLM inference:

  • Docker Desktop: Installed and running (Download here).
  • Hardware: Minimum 8GB RAM. (NVIDIA GPU recommended for speed, but CPU works for quantization).
  • Storage: At least 10GB free space for model weights.

🚀 Step-by-Step Deployment Guide

Step 1: Deploy the Ollama Backend

First, we pull the official image and run it in detached mode. This creates our API server.

docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama

Flag check: -v persists your downloaded models so you don't lose them when the container restarts.

Step 2: Install a Model (Llama3 or Phi3)

The container is empty by default. Let's send a command into the container to download a lightweight model like Phi-3.

docker exec -it ollama ollama pull phi3

Step 3: Deploy Open Web UI

Now for the interface. This command is longer because we must bridge the networking gap between the two containers.

docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

Success! Open your browser to http://localhost:3000. You will see the login screen (the first account created becomes the Admin).

🎥 Watch the Live Configuration

Prefer a visual walkthrough? Watch the complete setup process, including troubleshooting tips, in the video below.

🔧 Common Troubleshooting

1. "I can't connect to Ollama"
Ensure you used the --add-host flag in Step 3. If you are on Linux, you might need to use --network host instead.

2. Slow Performance?
Docker doesn't always use your GPU by default. You may need to install the NVIDIA Container Toolkit to pass GPU resources to the container.

© 2025 IOT Station. Privacy Policy

Comments

Post a Comment

Popular posts from this blog

How to Import CSV Files into PostgreSQL Automatically

-
Image
If you work with databases, manually uploading CSV data into PostgreSQL is a bottleneck. It’s slow, prone to human error, and creates a lag between data generation and availability. In this guide, I will show you how to build a Real-Time Data Pipeline . Using Python's watchdog library, we will create a script that "listens" to a specific folder. The moment a new CSV is dropped in, the script wakes up, parses the file, and pushes the data to PostgreSQL instantly. The Architecture Before writing code, it is helpful to understand the flow. We are moving from a manual "Pull" method to an automated "Event-Driven" method. The Trigger: The OS File System events (monitored by watchdog ). The Processor: A Python script using pandas or csv modules. The Destination: A PostgreSQL table using psycopg2 . Step 1: Prerequisites You will need a few libraries to handle the file monitoring and database connection....
Read more

Apache NiFi ETL Tutorial for Beginners | Installation & Data Pipeline Basics

-
Image
In the modern data landscape, the ability to move and transform information seamlessly is what separates a basic system from a professional data architecture. This process is known as ETL (Extract, Transform, Load) . Whether you are a student, an aspiring data engineer, or an IoT enthusiast, mastering ETL tools is a mandatory skill for building automated systems. In this comprehensive guide, I’m going to walk you through Apache NiFi . Unlike many other tools that require heavy coding, NiFi offers a powerful visual interface. I’ve spent a lot of time working with various data orchestrators, and NiFi remains one of my favorites due to its "drag-and-drop" simplicity combined with enterprise-grade power. Today, we will focus on getting it installed, secured, and running in a Windows environment. The Importance of Data Orchestration in IoT In my experience building IoT stations, the biggest challenge isn't just collecting data it's managing the flow. Imagine hav...
Read more

ESP32 TFT LCD 240x240 Tutorial-Display Images & DHT11 Sensor

-
Image
In this comprehensive tutorial, you will learn how to build a real-time temperature and humidity monitoring dashboard using an ESP32 , a 240×240 TFT LCD display (ST7789) , and a DHT11 sensor . The project demonstrates how to display text, icons, and live sensor data on a colorful TFT screen. This guide is designed for makers, students, and engineers who want to improve their skills in embedded systems , IoT development , and ESP32 graphical interfaces . 📺 Video Tutorial – Full Walkthrough This article complements the full step-by-step video tutorial below. In the video, you will see the complete wiring, TFT initialization, bitmap generation, and a live demo of temperature and humidity updates on the screen. 📌 Tip: Watching the video first will make the wiring and code much easier to follow. 🧰 Components Required ESP32 Development Board [Image of ESP32 development board pinout] 240×240 TFT LCD (ST7789 Driver) DHT11 Temperature & Humi...
Read more

Contact Form

Name

Email *

Message *