ant_ai.core.types

InvocationContext pydantic-model

Bases: BaseModel

Request-scoped execution context. Treat as read-only during a request.

Show JSON schema:

{
  "description": "Request-scoped execution context. Treat as read-only during a request.",
  "properties": {
    "session_id": {
      "title": "Session Id",
      "type": "string"
    },
    "llm_settings": {
      "anyOf": [
        {
          "additionalProperties": true,
          "type": "object"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "title": "Llm Settings"
    },
    "workflow_settings": {
      "anyOf": [
        {
          "additionalProperties": true,
          "type": "object"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "title": "Workflow Settings"
    }
  },
  "required": [
    "session_id"
  ],
  "title": "InvocationContext",
  "type": "object"
}

Config:

• frozen: True

Fields:

• session_id (str)
• llm_settings (dict[str, Any] | None)
• workflow_settings (dict[str, Any] | None)

Source code in src/ant_ai/core/types.py

class InvocationContext(BaseModel):
    """
    Request-scoped execution context. Treat as read-only during a request.
    """

    model_config = ConfigDict(frozen=True)

    session_id: str
    llm_settings: dict[str, Any] | None = Field(default=None)
    workflow_settings: dict[str, Any] | None = Field(default=None)

State pydantic-model

Bases: BaseModel

Shared mutable state passed through agent steps and workflow nodes.

Subclass to add domain-specific fields:

class MyState(State):
    user_id: str = ""

Show JSON schema:

{
  "$defs": {
    "Message": {
      "description": "Generic message used in a conversation",
      "properties": {
        "kind": {
          "const": "message",
          "default": "message",
          "title": "Kind",
          "type": "string"
        },
        "role": {
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "title": "Content"
        },
        "metadata": {
          "additionalProperties": true,
          "title": "Metadata",
          "type": "object"
        }
      },
      "required": [
        "role"
      ],
      "title": "Message",
      "type": "object"
    }
  },
  "description": "Shared mutable state passed through agent steps and workflow nodes.\n\nSubclass to add domain-specific fields:\n\n    class MyState(State):\n        user_id: str = \"\"",
  "properties": {
    "messages": {
      "items": {
        "$ref": "#/$defs/Message"
      },
      "title": "Messages",
      "type": "array"
    },
    "artefacts": {
      "items": {},
      "title": "Artefacts",
      "type": "array"
    }
  },
  "title": "State",
  "type": "object"
}

Fields:

• messages (list[Message])
• artefacts (list[Any])

Source code in src/ant_ai/core/types.py

class State(BaseModel):
    """Shared mutable state passed through agent steps and workflow nodes.

    Subclass to add domain-specific fields:

        class MyState(State):
            user_id: str = ""
    """

    messages: list[Message] = Field(default_factory=list)
    artefacts: list[Any] = Field(default_factory=list)

    @property
    def last_message(self) -> Message:
        """Returns the last message in the conversation, if any."""
        if not self.messages:
            raise ValueError("No messages in conversation")
        return self.messages[-1]

    def add_message(self, message: Message) -> None:
        self.messages.append(message)

last_message property

last_message: Message

Returns the last message in the conversation, if any.