Response Types¶

Response types define structured output formats for agents. They can be defined directly in YAML or imported from Python code.

Type Safety

While YAML configuration is convenient, defining response types as Pydantic models in Python code provides better type safety, IDE support, and reusability:

from pydantic import BaseModel

class AnalysisResult(BaseModel):
    success: bool
    issues: list[str]
    severity: str

Response Configuration¶

Each response definition includes:

responses:
  MyResponse:
    response_schema:  # Schema definition (required)
      type: "inline"  # or "import"
      # schema details...
    description: "Optional description of the response"
    result_tool_name: "final_result"  # Tool name for result creation
    result_tool_description: "Create the final result"  # Tool description
    result_retries: 3  # Number of validation retries

Inline Responses¶

Define response structure directly in YAML:

Simple Status Response¶

responses:
  StatusResponse:
    response_schema:
      type: "inline"
      description: "Simple operation result with status"
      fields:
        success:
          type: "bool"
          description: "Whether operation succeeded"
        message:
          type: "str"
          description: "Status message or error details"

Analysis Result¶

responses:
  CodeAnalysis:
    response_schema:
      type: "inline"
      description: "Code analysis results with issues"
      fields:
        issues:
          type: "list[str]"
          description: "List of found issues"
        severity:
          type: "str"
          description: "Overall severity level"
        locations:
          type: "list[str]"
          description: "Source code locations"
    result_tool_name: "create_analysis"
    result_tool_description: "Create code analysis result"

Complex Response¶

responses:
  DataProcessingResult:
    response_schema:
      type: "inline"
      description: "Complex data processing result"
      fields:
        success:
          type: "bool"
          description: "Operation success"
        records_processed:
          type: "int"
          description: "Number of processed records"
        errors:
          type: "list[str]"
          description: "List of errors if any"
        metrics:
          type: "dict[str, float]"
          description: "Processing metrics"

Imported Responses¶

Import response types from Python code:

Python Type Import¶

responses:
  AdvancedAnalysis:
    response_schema:
      type: "import"
      import_path: "myapp.types:AnalysisResult"

Package Response Type¶

responses:
  MetricsResult:
    response_schema:
      type: "import"
      import_path: "myapp.analysis:MetricsResponse"

Using Response Types¶

Assign to Agent¶

agents:
  analyzer:
    model: "openai:gpt-4"
    result_type: "CodeAnalysis"  # Reference response by name

Inline with Custom Tool Name¶

agents:
  processor:
    result_type:
      response_schema:
        type: "inline"  # Direct inline definition
        fields:
          success:
            type: "bool"
          details:
            type: "str"
      result_tool_name: "create_result"  # Custom tool name
      result_tool_description: "Create the final analysis result"
      result_retries: 2  # Number of validation attempts

Available Field Types¶

str: Text strings
int: Integer numbers
float: Floating point numbers
bool: Boolean values
list[type]: Lists of values (e.g., list[str], list[int])
dict[key_type, value_type]: Dictionaries
datetime: Date and time values
Custom types through imports

Schema Definition¶

The response schema defines the structure of the return value:

response_schema:
  type: "inline"  # Define structure inline
  description: "Result description"
  fields:
    # Field definitions

# OR

response_schema:
  type: "import"  # Import from Python
  import_path: "myapp.types:MyResponseType"