Configuration FAQ

What is configuration?

The configuration system manages settings for Attune AI agents, workflows, and integrations. It handles both file-based configuration and environment variables, with support for different model providers and runtime environments.

When should I use it?

Use the configuration system when you need to:

• Set up agent parameters (model, temperature, token limits)
• Configure workflow execution modes
• Manage environment-specific settings
• Switch between different LLM providers
• Store persistent configuration for book production or other workflows

How do I load configuration?

Use load_unified_config() to load configuration from the default locations, or ConfigLoader for custom paths:

from attune.config import load_unified_config

# Load from default paths
config = load_unified_config()

# Or specify a custom path
config = load_unified_config("./my-config.json")

How do I work with environment variables?

Use get_attune_env() to read environment variables with fallback support:

from attune.config import get_attune_env

# Checks ATTUNE_MODEL first, then EMPATHY_MODEL
model = get_attune_env("MODEL", default="gpt-4")

What configuration formats are supported?

The system supports JSON configuration files. It searches these default locations:

• ./attune.config.json
• ~/.attune/config.json
• ~/.config/attune/config.json

How do I configure different agents?

Use UnifiedAgentConfig for general agents or BookProductionConfig for book production workflows:

# General agent configuration
agent_config = config.for_book_production()

# Access model settings
model_id = agent_config.get_model_id()

How do I debug configuration issues?

Run the configuration tests first:

pytest -k "config" -v

If tests pass but your code fails, check that:

• Your configuration file exists and is valid JSON
• Environment variables use the correct ATTUNE_ prefix
• Required configuration sections are present

Where are the source files?

• src/attune/config/**

Tags: config, settings