LLM provider

To support the many Fabricate features that rely on AI, you need to configure a connection to either:

• Amazon Bedrock
• Anthropic
• Azure AI Foundry
• Google Gemini Enterprise Agent Platform (formerly Vertex AI)

Amazon Bedrock

To use Amazon Bedrock as your Claude LLM provider, in .env, uncomment and set the following:

FABRICATE_BEDROCK_REGION=   # Bedrock model region prefix to prepend to model IDs.
                            # For example, us.anthropic.claude-<model>.
                            # Default is global, to enable cross-region inference.
FABRICATE_BEDROCK_API_KEY=  # API key for Amazon Bedrock
FABRICATE_AWS_REGION= # AWS Region where Fabricate runs.

Note that FABRICATE_AWS_REGION is also used if you use Amazon S3 for cloud storage.

Anthropic

To use Anthropic as your Claude LLM provider, in .env, uncomment and set the following:

FABRICATE_ANTHROPIC_BASE_URL=  # URL to connect to Anthropic
FABRICATE_ANTHROPIC_API_KEY=   # API key for Anthropic

Azure AI Foundry

To use Azure AI Foundry as your Claude LLM provider, in .env, uncomment and set the following:

FABRICATE_AZURE_URI=        # Base URL for Azure AI Foundry Anthropic-compatible endpoint.
FABRICATE_AZURE_API_KEY=    # API key for the Azure AI Foundry endpoint.

Google Gemini Enterprise Agent Platform

To use Google Gemini Enterprise Agent Platform (formerly Vertex AI) as your Claude LLM provider, in .env, uncomment and set the following:

FABRICATE_VERTEXAI_PROJECT_ID=  # Google Cloud project ID (required).
                                # Falls back to GOOGLE_CLOUD_PROJECT.
FABRICATE_VERTEXAI_GOOGLE_CLOUD_LOCATION=  # Vertex AI region (required).
                                           # Falls back to GOOGLE_CLOUD_LOCATION.
FABRICATE_VERTEXAI_SERVICE_ACCOUNT_KEY=  # Service account JSON key.
                                         # Falls back to VERTEXAI_SERVICE_ACCOUNT_KEY.