AI Configuration

Dockerizer can use AI to generate Docker configurations when automatic detection fails or for unsupported frameworks. This page explains how to configure different AI providers.

When Is AI Used?

AI generation is triggered in these scenarios:

Low confidence detection - When framework detection confidence is below 70%
Unknown frameworks - Languages or frameworks without built-in templates
Forced AI mode - When using the --ai flag
Complex setups - Monorepos or multi-service architectures

Supported Providers

Anthropic Claude Recommended

Claude models provide excellent Docker configuration generation with deep understanding of containerization best practices.

Setup

export ANTHROPIC_API_KEY="sk-ant-api03-..."

Usage

# Use Claude (default model)
dockerizer --ai --ai-provider anthropic

# Specify model
dockerizer --ai --ai-provider anthropic --ai-model claude-sonnet-4-20250514

Model	Best For
`claude-sonnet-4-20250514`	Balanced speed and quality (default)
`claude-opus-4-20250514`	Complex configurations, best quality
`claude-haiku-3-20250514`	Fast generation, simple projects

OpenAI GPT

OpenAI models are widely used and provide reliable Docker configuration generation.

Setup

export OPENAI_API_KEY="sk-proj-..."

Usage

# Use OpenAI (default model)
dockerizer --ai --ai-provider openai

# Specify model
dockerizer --ai --ai-provider openai --ai-model gpt-4o

Model	Best For
`gpt-4o`	Best quality, multimodal (default)
`gpt-4o-mini`	Faster, cost-effective
`o1`	Advanced reasoning for complex setups

Ollama Local

Run AI locally without sending code to external services. Great for privacy-sensitive projects.

Setup

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Pull a model
ollama pull llama3.2

# Ollama runs on localhost by default
# Optional: set custom host
export OLLAMA_HOST="http://localhost:11434"

Usage

# Use Ollama with default model
dockerizer --ai --ai-provider ollama

# Specify model
dockerizer --ai --ai-provider ollama --ai-model codellama

Model	Best For
`llama3.2`	General purpose (default)
`codellama`	Code-focused tasks
`deepseek-coder`	Specialized coding model

Environment Variables

Configure AI settings via environment variables:

Variable	Description	Example
`ANTHROPIC_API_KEY`	Anthropic API key	`sk-ant-api03-...`
`OPENAI_API_KEY`	OpenAI API key	`sk-proj-...`
`OLLAMA_HOST`	Ollama server URL	`http://localhost:11434`
`DOCKERIZER_AI_PROVIDER`	Default provider	`anthropic`
`DOCKERIZER_AI_MODEL`	Default model	`claude-sonnet-4-20250514`

Configuration File

Create a .dockerizer.yml in your home directory or project root:

# ~/.dockerizer.yml
ai:
  provider: anthropic
  model: claude-sonnet-4-20250514

# Or for OpenAI
ai:
  provider: openai
  model: gpt-4o

# Or for local Ollama
ai:
  provider: ollama
  host: http://localhost:11434
  model: llama3.2

Provider Priority

Dockerizer checks for AI providers in this order:

Command-line flags (--ai-provider)
Environment variable (DOCKERIZER_AI_PROVIDER)
Configuration file (.dockerizer.yml)
Auto-detect based on available API keys

Best Practices

Security Tip: Never commit API keys to version control. Use environment variables or a .env file (added to .gitignore).

Start with detection - Let Dockerizer try automatic detection first. Only use AI when needed.
Review generated configs - Always review AI-generated Dockerfiles before using in production.
Use local models for sensitive code - If your code is proprietary, consider Ollama for local processing.
Set defaults - Configure your preferred provider in ~/.dockerizer.yml to avoid repeated flags.

Troubleshooting

API Key Issues

# Verify API key is set
echo $ANTHROPIC_API_KEY
echo $OPENAI_API_KEY

# Test with verbose output
dockerizer --ai --verbose

Ollama Connection Issues

# Check if Ollama is running
curl http://localhost:11434/api/tags

# Restart Ollama
ollama serve

Model Not Found

# List available Ollama models
ollama list

# Pull missing model
ollama pull llama3.2