Docker with Crew Template README file

This is a template for getting CrewAI up and running in a Dockerfile. Author: Vinod Ralh

1. Setup Python environment
2. Setup Dockerfile
3. Build and run CrewAI project using Docker commands

This uses default instructions from: https://docs.crewai.com/installation

It was inspired after watching "The Surprising Power of CrewAI in Docker" https://youtu.be/JGID_du9-So?si=Dm_XE1vfHJOVQH3d. Thanks @TylerReedAI

ToDo: Replay, test, train arguments to crewai don't work

Setup Environment

Tested on Mac Silicone

Note after you clone this repo, you need to add a .env file If you follow these instruction you should be able to create your own clean version.

Setup Python

Install specific version of python

You may need to install pyenv, which you can via homebrew brew install pyenv

List available Python versions pyenv install --list

Install the desired Python version pyenv install 3.11.9

Set the Python version globally or locally: pyenv local 3.11.9 pyenv global 3.11.9

Verify the installation: python --version

Setup CrewAI

Installing CrewAI pip install 'crewai[tools] pip install crewai crewai-tools

Verify Installation pip freeze | grep crewai

Create a requirements.txt file pip freeze > requirements.txt

Create Crew project crewai create crew <project_name> crewai create crew docker_crew_template

Lock and install the dependencies cd <project_name> cd docker_crew_template crewai install

Run Crew crewai run

Setup Docker

Note that this Dockerfile was generated using OpenAI ChatGPT4o and then modified.

Dockerfile
FROM python:3.11.9
# Set the working directory in the container
WORKDIR /app
# Copy requirements and install dependencies
COPY requirements.txt .
RUN pip install -r requirements.txt
# Set the PYTHONPATH to include src/
ENV PYTHONPATH=/app/src
# Copy the application code
COPY . .
# Expose the app port
EXPOSE 5000
# Specify the command to run on container start
CMD ["crewai", "run"]

Dockerfile

Run CrewAI Dockerfile

Stop and delete the container before rebuilding it docker stop my-app-crew_container 2>/dev/null docker rm my-app-crew_container 2>/dev/null

Clean up all stopped containers in one step docker container prune -f

Build the Docker Image docker build -t my-crew_app .

Run the Docker Container docker run -d -p 5000:5000 --name my-app-crew_container my-crew_app

My port 5000 is in use - so using another one 5001 docker run -d -p 5001:5000 --name my-app-crew_container my-crew_app

View logs

docker ps -a docker logs <container-name-or-id> docker logs my-app-crew_container curl http://localhost:5001

Run application with arguments in Docker

Note: Arguments such as test, train, replay aren't working as yet!

Run the Default Behavior (run): docker run my-crew_app

Run a Specific Command: docker run my-crew_app train docker run my-crew_app replay 12345

Testing Locally (without Docker): python src/docker_crew_template/main.py train

Debugging in Docker - you can run the container interactively: docker run -it my-crew_app /bin/bash

Once inside the container: python src/docker_crew_template/main.py