An LLM-powered tool that automatically containerizes your project files without human involvement. It uses an LLM API to read your code, generate an optimized Dockerfile with multi-stage builds and Alpine/slim base images, and then builds and validates the Docker image — all from a single command.

autoDocker is a smart build wrapper that:

1. Extracts project intent — reads your source files to understand the project structure (WorkspaceManager)
2. Drafts an optimized Dockerfile — uses an LLM to generate a secure, minimal Dockerfile (LLMArchitect)
3. Builds the Docker image — communicates with Docker on your machine (DockerBuilder)
4. Self-heals on errors — if the build fails, it automatically asks the LLM to fix the Dockerfile and retries
5. Validates the result — runs the container briefly to confirm it starts correctly (Validator)

• Python 3.10+
• Docker — install Docker Desktop and make sure it is running
• A Groq API key — free tier available at https://console.groq.com

# Clone the repository git clone https://github.com/Qrytics/autoDocker.git cd autoDocker # Install the package (creates the `autodocker` CLI command) pip install -e .

Export your API key before running:

export GROQ_API_KEY='your_key_here'

Using a different LLM provider? autoDocker uses LiteLLM under the hood, so any provider it supports works. For example, to use Gemini set GEMINI_API_KEY and pass --model gemini/gemini-1.5-flash.

This example containerizes a simple "Hello, World!" Python script in under a minute.

1. Create a sample project zip:

mkdir hello_world echo 'print("Hello, World!")' > hello_world/main.py zip -r hello_world.zip hello_world/

2. Run autoDocker:

autodocker hello_world.zip --tag hello-world:v1

You should see autoDocker unpack the project, generate a Dockerfile, build the image, and confirm it runs successfully.

3. Run your container manually (optional):

docker run --rm hello-world:v1 # Hello, World!

autodocker <source> [options] 

# Basic usage with a zip file autodocker my_project.zip # Set a custom image tag autodocker my_project.zip --tag myapp:v1.0 # Containerize a public GitHub repository autodocker https://github.com/user/repo --tag myrepo:latest # Use a different LLM model and skip the runtime test autodocker my_project.zip --model gemini/gemini-1.5-flash --skip-test # Show all options autodocker --help