Copilot Agent 365 - Docker Deployment Tutorial

1

Overview

Copilot Agent 365 is an enterprise-grade AI assistant that runs entirely on your local machine using Ollama and Llama 3.1. No API keys required, no data leaves your system, and it's completely free to run.

✨

Key Features

🦙 Llama 3.1 - State-of-the-art open source LLM
🔒 100% Local - Your data never leaves your machine
🐳 Docker Ready - One command deployment
🧩 Modular Agents - Extensible architecture
💾 Persistent Memory - Remembers context across sessions
☁️ Azure OpenAI - Optional cloud integration

System Architecture

You

Browser UI

↓

Docker

Port 7071

↓

Agent App

Python/Flask

Ollama

Llama 3.1

2

Prerequisites

Before you begin, make sure you have the following installed on your system.

Docker Desktop

Required for running containers. Download here
8GB+ RAM

Llama 3.1 requires at least 8GB of memory to run smoothly
10GB+ Disk Space

For Docker images and the Llama 3.1 model weights
Git (Optional)

For cloning the repository. You can also download the ZIP.

💡

GPU Acceleration: If you have an NVIDIA GPU, you'll get much faster responses. Make sure you have the NVIDIA Container Toolkit installed.

3

Quick Start

Get up and running in under 5 minutes with these simple steps.

Clone the Repository

Download the project files from GitHub:

bash

git clone https://github.com/kody-w/copilot-agent-365-docker.git
cd copilot-agent-365-docker

Start with Docker Compose

Choose the right command for your system:

🖥️

CPU Only - For systems without NVIDIA GPU

bash

docker compose -f docker-compose.cpu.yml up -d

🎮

With GPU - For NVIDIA GPU acceleration

bash

docker compose up -d

Wait for Model Download

The first run will download the Llama 3.1 model (~4.7GB). Monitor progress:

bash

docker logs -f ollama

Wait until you see "Model ready!" in the logs.

Access the Chat UI

Open your browser and navigate to:

🚀 Open http://localhost:7071

Terminal

$ docker compose -f docker-compose.cpu.yml up -d

[+] Running 2/2

✔ Container ollama Started

✔ Container agent_function Started

$ docker logs ollama

Waiting for Ollama server to be ready...

Ollama server ready. Pulling llama3.1 model...

Model ready!

4

Configuration

Customize your deployment with environment variables.

Variable	Default	Description
`USE_OLLAMA`	true	Use local Ollama instead of Azure OpenAI
`OLLAMA_MODEL_NAME`	llama3.1	Which Ollama model to use
`ASSISTANT_NAME`	CopilotAgent365	Display name for the assistant
`FUNCTION_APP_PORT`	7071	Port for the web interface

☁️

Azure OpenAI Mode

To use Azure OpenAI instead of local Ollama:

bash

docker run -p 7071:7071 \
  -e USE_OLLAMA=false \
  -e AZURE_OPENAI_API_KEY=your-key \
  -e AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com/ \
  -e AZURE_OPENAI_DEPLOYMENT_NAME=gpt-4o \
  kodywf/copilot-agent-365:latest

5

Architecture Deep Dive

Understanding how the components work together.

🐳

Docker Services

agent_function - The main application container running the Python Flask app with Azure Functions runtime.

ollama - The Ollama container that serves the Llama 3.1 model via an OpenAI-compatible API.

🧩

Agent System

The modular agent architecture allows for extensible functionality:

ContextMemoryAgent - Recalls conversation history
ManageMemoryAgent - Stores facts, preferences, and insights
ScriptedDemoAgent - Interactive demos and walkthroughs
GithubAgentLibraryManager - Manages agent installations

6

Docker Hub

Pull the pre-built image directly from Docker Hub.

bash

docker pull kodywf/copilot-agent-365:latest

🏷️

Available Tags

latest - Most recent stable build
1.0.0 - Initial release with Llama 3.1 support

🐳 View on Docker Hub 📦 View on GitHub

7

Troubleshooting

Common issues and how to resolve them.

❌

Container won't start

Check if Docker is running and ports aren't in use:

bash

docker ps -a
lsof -i :7071
lsof -i :11434

🐢

Slow responses

Without GPU acceleration, responses may take 30-60 seconds. Consider:

Using a smaller model: OLLAMA_MODEL_NAME=llama3.1:8b
Adding more RAM to Docker Desktop
Using GPU acceleration if available

🔄

Reset everything

To completely reset and start fresh:

bash

docker compose down -v
docker system prune -a
docker compose -f docker-compose.cpu.yml up -d

⚠️

Need more help? Open an issue on GitHub