config-example.yaml

general:
  trace_content_enabled: true # Optional, defaults to true, set to false to disable tracing of request and response content
providers:
  # Azure OpenAI configuration
  - key: azure-openai
    type: azure
    api_key: "<your-azure-api-key>"
    resource_name: "<your-resource-name>"
    api_version: "<your-api-version>"

  # OpenAI configuration
  - key: openai
    type: openai
    api_key: "<your-openai-api-key>"
    # or use environment variable
    api_key: ${OPENAI_API_KEY}
    base_url: "optional base url. If not provided, defaults to https://api.openai.com/v1"

  # Anthropic configuration
  - key: anthropic
    type: anthropic
    api_key: "<your-anthropic-api-key>"

  - key: bedrock
    type: bedrock
    api_key: ""# Not used for AWS Bedrock
    region: "<your-aws-region>" # like "us-east-1"
    inference_profile_id: "<your-inference-profile-id>" # like "us"
    # Authentication options - choose one:
    # Option 1: Use IAM roles (recommended for production)
    use_iam_role: true  # Use IAM roles for service accounts (IRSA) or instance profiles
    # Option 2: Use explicit credentials (for development/testing)
    # use_iam_role: "false"  # or omit this line
    # AWS_ACCESS_KEY_ID: "<your-aws-access-key>"
    # AWS_SECRET_ACCESS_KEY: "<your-aws-secret-key>"
    # AWS_SESSION_TOKEN: "<your-session-token>"  # Optional

  # Vertex AI configuration
  # Uses service account authentication
  - key: vertexai
    type: vertexai
    api_key: ""  # Required field but not used with service account auth
    project_id: "<your-gcp-project-id>"
    location: "<your-gcp-region>"  # e.g., us-central1
    credentials_path: "/path/to/service-account.json"  # Path to your service account key file

models:
  # OpenAI Models
  - key: gpt-4
    type: gpt-4
    provider: openai
  - key: gpt-3.5-turbo
    type: gpt-3.5-turbo
    provider: openai

  # Anthropic Models
  - key: claude-3-5-sonnet
    type: claude-3-5-sonnet-20241022
    provider: anthropic
  - key: claude-3-haiku
    type: claude-3-haiku-20240307
    provider: anthropic

  # Azure OpenAI Models
  - key: gpt-4-azure
    type: gpt-4
    provider: azure-openai
    deployment: "<your-deployment-name>"
  - key: gpt-35-turbo-azure
    type: gpt-35-turbo
    provider: azure-openai
    deployment: "<your-deployment-name>"

  # Bedrock Models
  - key: bedrock-model
    # some models are region specific, it is a good idea to get ARN from cross region reference tab
    type: "< model-id or Inference profile ARN or Inference profile ID>"
    provider: bedrock
    model_provider: "anthropic" # can be: ai21, titan, anthropic
    model_version: "v2:0" # optional, defaults to "v1:0"

  # Vertex AI Models
  # Chat and Completion model
  - key: gemini-1.5-flash
    type: gemini-1.5-flash  # Supports both chat and completion endpoints
    provider: vertexai
  # Embeddings model
  - key: textembedding-gecko
    type: textembedding-gecko  # Supports embeddings endpoint
    provider: vertexai
    deployment: "<your-deployment>"

pipelines:
  # Default pipeline for chat completions
  - name: default
    type: chat
    plugins:
      - logging:
          level: info  # Supported levels: debug, info, warning, error
      - tracing:  # Optional tracing configuration
          endpoint: "https://api.traceloop.com/v1/traces"
          api_key: "<your-traceloop-api-key>"
      - model-router:
          models:  # List the models you want to use for chat
            - gpt-4
            - gpt-4-azure
            - claude-3-5-sonnet
            - gemini-1.5-flash

  # Pipeline for text completions
  - name: completions
    type: completion
    plugins:
      - model-router:
          models:  # List the models you want to use for completions
            - gpt-3.5-turbo
            - gpt-35-turbo-azure
            - claude-3-haiku
            - gemini-1.5-flash

  # Pipeline for embeddings
  - name: embeddings
    type: embeddings
    plugins:
      - model-router:
          models:  # List the models you want to use for embeddings
            - textembedding-gecko