Skip to content

LangGraph Integration

Execute LangGraph StateGraphs securely through Sequrity's Dual-LLM runtime with policy enforcement.

Requirements

  • LangGraph must be installed: pip install langgraph
  • Requires Dual LLM mode (features.agent_arch == "dual-llm")
  • strip_response_content must be False in fine-grained config

langgraph

LangGraph resource — client.control.langgraph.

Classes:

LangGraphResource 

LangGraphResource(transport: ControlSyncTransport)

LangGraph execution — client.control.langgraph.

Methods:

  • run –

    Compile and execute a LangGraph StateGraph with Sequrity security.

Source code in src/sequrity/control/resources/langgraph/__init__.py 
20
21
def __init__(self, transport: ControlSyncTransport) -> None:
    self._transport = transport

run 

run(
    graph: StateGraph,
    initial_state: dict,
    model: str,
    *,
    initial_state_meta: MetaData | None = None,
    max_exec_steps: int = 20,
    node_functions: dict[str, Callable] | None = None,
    internal_node_mapping: dict[str, str] | None = None,
    timeout: float = 300.0,
    llm_api_key: str | None | _NotGiven = NOT_GIVEN,
    features: FeaturesHeader | None | _NotGiven = NOT_GIVEN,
    security_policy: SecurityPolicyHeader | None | _NotGiven = NOT_GIVEN,
    fine_grained_config: FineGrainedConfigHeader | None | _NotGiven = NOT_GIVEN,
    provider: LlmServiceProvider | LlmServiceProviderStr | None | _NotGiven = NOT_GIVEN,
    custom_headers: dict[str, str] | None = None,
) -> dict

Compile and execute a LangGraph StateGraph with Sequrity security.

Parameters:

  • graph

    (StateGraph) –

    A LangGraph StateGraph instance defining the agent workflow.

  • initial_state

    (dict) –

    Initial state dictionary to start graph execution.

  • model

    (str) –

    The LLM model identifier for agent nodes.

  • initial_state_meta

    (MetaData | None, default: None ) –

    Optional metadata for the initial state.

  • max_exec_steps

    (int, default: 20 ) –

    Maximum number of execution steps. Defaults to 20.

  • node_functions

    (dict[str, Callable] | None, default: None ) –

    Custom node function implementations.

  • internal_node_mapping

    (dict[str, str] | None, default: None ) –

    Mapping for internal node names.

  • timeout

    (float, default: 300.0 ) –

    Timeout in seconds for each API request. Defaults to 300.0.

  • llm_api_key

    (str | None | _NotGiven, default: NOT_GIVEN ) –

    LLM provider API key override.

  • features

    (FeaturesHeader | None | _NotGiven, default: NOT_GIVEN ) –

    Security features override.

  • security_policy

    (SecurityPolicyHeader | None | _NotGiven, default: NOT_GIVEN ) –

    Security policy override.

  • fine_grained_config

    (FineGrainedConfigHeader | None | _NotGiven, default: NOT_GIVEN ) –

    Fine-grained config override.

  • provider

    (LlmServiceProvider | LlmServiceProviderStr | None | _NotGiven, default: NOT_GIVEN ) –

    LLM service provider override.

Returns:

  • dict –

    Dictionary containing the final state and execution results.

Source code in src/sequrity/control/resources/langgraph/__init__.py 
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
def run(
    self,
    graph: StateGraph,
    initial_state: dict,
    model: str,
    *,
    initial_state_meta: MetaData | None = None,
    max_exec_steps: int = 20,
    node_functions: dict[str, Callable] | None = None,
    internal_node_mapping: dict[str, str] | None = None,
    timeout: float = 300.0,
    # Sequrity overrides
    llm_api_key: str | None | _NotGiven = NOT_GIVEN,
    features: FeaturesHeader | None | _NotGiven = NOT_GIVEN,
    security_policy: SecurityPolicyHeader | None | _NotGiven = NOT_GIVEN,
    fine_grained_config: FineGrainedConfigHeader | None | _NotGiven = NOT_GIVEN,
    provider: LlmServiceProvider | LlmServiceProviderStr | None | _NotGiven = NOT_GIVEN,
    custom_headers: dict[str, str] | None = None,
) -> dict:
    """Compile and execute a LangGraph StateGraph with Sequrity security.

    Args:
        graph: A LangGraph StateGraph instance defining the agent workflow.
        initial_state: Initial state dictionary to start graph execution.
        model: The LLM model identifier for agent nodes.
        initial_state_meta: Optional metadata for the initial state.
        max_exec_steps: Maximum number of execution steps. Defaults to 20.
        node_functions: Custom node function implementations.
        internal_node_mapping: Mapping for internal node names.
        timeout: Timeout in seconds for each API request. Defaults to 300.0.
        llm_api_key: LLM provider API key override.
        features: Security features override.
        security_policy: Security policy override.
        fine_grained_config: Fine-grained config override.
        provider: LLM service provider override.

    Returns:
        Dictionary containing the final state and execution results.
    """
    from ._runner import run_graph_sync

    return run_graph_sync(
        transport=self._transport,
        model=model,
        graph=graph,
        initial_state=initial_state,
        initial_state_meta=initial_state_meta,
        max_exec_steps=max_exec_steps,
        node_functions=node_functions,
        internal_node_mapping=internal_node_mapping,
        timeout=timeout,
        llm_api_key=llm_api_key,
        features=features,
        security_policy=security_policy,
        fine_grained_config=fine_grained_config,
        provider=provider,
        custom_headers=custom_headers,
    )

LangGraphChatCompletionRequest pydantic-model

Extended chat completion request for LangGraph execution (OpenAI-compatible providers).

Show JSON schema: 
{
  "$defs": {
    "AssistantMessage": {
      "description": "Assistant message.",
      "properties": {
        "role": {
          "const": "assistant",
          "description": "The role of the messages author, in this case 'assistant'.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "items": {
                "anyOf": [
                  {
                    "$ref": "#/$defs/ContentPartText"
                  },
                  {
                    "$ref": "#/$defs/ContentPartRefusal"
                  }
                ]
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The contents of the assistant message. Required unless tool_calls or function_call is specified.",
          "title": "Content"
        },
        "name": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "An optional name for the participant. Provides the model information to differentiate between participants of the same role.",
          "title": "Name"
        },
        "refusal": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The refusal message by the assistant.",
          "title": "Refusal"
        },
        "audio": {
          "anyOf": [
            {
              "$ref": "#/$defs/AssistantMessageAudio"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Data about a previous audio response from the model."
        },
        "tool_calls": {
          "anyOf": [
            {
              "items": {
                "$ref": "#/$defs/ToolCall"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The tool calls generated by the model, such as function calls.",
          "title": "Tool Calls"
        },
        "function_call": {
          "anyOf": [
            {
              "$ref": "#/$defs/FunctionCallResult"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Deprecated and replaced by tool_calls. The name and arguments of a function that should be called."
        }
      },
      "required": [
        "role"
      ],
      "title": "AssistantMessage",
      "type": "object"
    },
    "AssistantMessageAudio": {
      "description": "Audio reference in assistant message.",
      "properties": {
        "id": {
          "description": "Unique identifier for a previous audio response from the model.",
          "title": "Id",
          "type": "string"
        }
      },
      "required": [
        "id"
      ],
      "title": "AssistantMessageAudio",
      "type": "object"
    },
    "ContentPartFile": {
      "description": "File content part.",
      "properties": {
        "type": {
          "const": "file",
          "description": "The type of the content part. Always 'file'.",
          "title": "Type",
          "type": "string"
        },
        "file": {
          "$ref": "#/$defs/FileData",
          "description": "The file data."
        }
      },
      "required": [
        "type",
        "file"
      ],
      "title": "ContentPartFile",
      "type": "object"
    },
    "ContentPartImage": {
      "description": "Image content part.",
      "properties": {
        "type": {
          "const": "image_url",
          "description": "The type of the content part.",
          "title": "Type",
          "type": "string"
        },
        "image_url": {
          "$ref": "#/$defs/ImageUrl",
          "description": "The image URL details."
        }
      },
      "required": [
        "type",
        "image_url"
      ],
      "title": "ContentPartImage",
      "type": "object"
    },
    "ContentPartInputAudio": {
      "description": "Input audio content part.",
      "properties": {
        "type": {
          "const": "input_audio",
          "description": "The type of the content part.",
          "title": "Type",
          "type": "string"
        },
        "input_audio": {
          "$ref": "#/$defs/InputAudioData",
          "description": "The input audio data."
        }
      },
      "required": [
        "type",
        "input_audio"
      ],
      "title": "ContentPartInputAudio",
      "type": "object"
    },
    "ContentPartRefusal": {
      "description": "Refusal content part.",
      "properties": {
        "type": {
          "const": "refusal",
          "description": "The type of the content part.",
          "title": "Type",
          "type": "string"
        },
        "refusal": {
          "description": "The refusal message.",
          "title": "Refusal",
          "type": "string"
        }
      },
      "required": [
        "type",
        "refusal"
      ],
      "title": "ContentPartRefusal",
      "type": "object"
    },
    "ContentPartText": {
      "description": "Text content part.",
      "properties": {
        "type": {
          "const": "text",
          "description": "The type of the content part.",
          "title": "Type",
          "type": "string"
        },
        "text": {
          "description": "The text content.",
          "title": "Text",
          "type": "string"
        }
      },
      "required": [
        "type",
        "text"
      ],
      "title": "ContentPartText",
      "type": "object"
    },
    "DeveloperMessage": {
      "description": "Developer message.",
      "properties": {
        "role": {
          "const": "developer",
          "description": "The role of the messages author, in this case 'developer'.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "items": {
                "$ref": "#/$defs/ContentPartText"
              },
              "type": "array"
            }
          ],
          "description": "The contents of the developer message.",
          "title": "Content"
        },
        "name": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "An optional name for the participant. Provides the model information to differentiate between participants of the same role.",
          "title": "Name"
        }
      },
      "required": [
        "role",
        "content"
      ],
      "title": "DeveloperMessage",
      "type": "object"
    },
    "FileData": {
      "description": "File data.",
      "properties": {
        "file_data": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The base64 encoded file data, used when passing the file to the model as a string.",
          "title": "File Data"
        },
        "file_id": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The ID of an uploaded file to use as input.",
          "title": "File Id"
        },
        "filename": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The name of the file, used when passing the file to the model as a string.",
          "title": "Filename"
        }
      },
      "title": "FileData",
      "type": "object"
    },
    "FunctionCallResult": {
      "description": "Function call result in assistant message.",
      "properties": {
        "arguments": {
          "description": "The arguments to call the function with, as generated by the model in JSON format.",
          "title": "Arguments",
          "type": "string"
        },
        "name": {
          "description": "The name of the function to call.",
          "title": "Name",
          "type": "string"
        }
      },
      "required": [
        "arguments",
        "name"
      ],
      "title": "FunctionCallResult",
      "type": "object"
    },
    "FunctionDefinition": {
      "description": "Function definition for tools.",
      "properties": {
        "name": {
          "description": "The name of the function to be called. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64.",
          "title": "Name",
          "type": "string"
        },
        "description": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "A description of what the function does, used by the model to choose when and how to call the function.",
          "title": "Description"
        },
        "parameters": {
          "anyOf": [
            {
              "additionalProperties": true,
              "type": "object"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The parameters the functions accepts, described as a JSON Schema object.",
          "title": "Parameters"
        },
        "strict": {
          "anyOf": [
            {
              "type": "boolean"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Whether to enable strict schema adherence when generating the function call.",
          "title": "Strict"
        }
      },
      "required": [
        "name"
      ],
      "title": "FunctionDefinition",
      "type": "object"
    },
    "FunctionMessage": {
      "description": "Function message (deprecated).",
      "properties": {
        "role": {
          "const": "function",
          "description": "The role of the messages author, in this case 'function'.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "description": "The contents of the function message.",
          "title": "Content"
        },
        "name": {
          "description": "The name of the function.",
          "title": "Name",
          "type": "string"
        }
      },
      "required": [
        "role",
        "content",
        "name"
      ],
      "title": "FunctionMessage",
      "type": "object"
    },
    "FunctionTool": {
      "description": "Function tool.",
      "properties": {
        "type": {
          "const": "function",
          "description": "The type of the tool. Currently, only 'function' is supported.",
          "title": "Type",
          "type": "string"
        },
        "function": {
          "$ref": "#/$defs/FunctionDefinition",
          "description": "The function definition."
        }
      },
      "required": [
        "type",
        "function"
      ],
      "title": "FunctionTool",
      "type": "object"
    },
    "ImageUrl": {
      "description": "Image URL details.",
      "properties": {
        "url": {
          "description": "Either a URL of the image or the base64 encoded image data.",
          "title": "Url",
          "type": "string"
        },
        "detail": {
          "anyOf": [
            {
              "enum": [
                "auto",
                "low",
                "high"
              ],
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Specifies the detail level of the image. 'low', 'high', or 'auto' for automatic selection.",
          "title": "Detail"
        }
      },
      "required": [
        "url"
      ],
      "title": "ImageUrl",
      "type": "object"
    },
    "InputAudioData": {
      "description": "Input audio data.",
      "properties": {
        "data": {
          "description": "Base64 encoded audio data.",
          "title": "Data",
          "type": "string"
        },
        "format": {
          "description": "The format of the audio data.",
          "enum": [
            "wav",
            "mp3"
          ],
          "title": "Format",
          "type": "string"
        }
      },
      "required": [
        "data",
        "format"
      ],
      "title": "InputAudioData",
      "type": "object"
    },
    "JsonSchema": {
      "additionalProperties": false,
      "description": "JSON schema definition.",
      "properties": {
        "name": {
          "description": "The name of the response format. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64.",
          "title": "Name",
          "type": "string"
        },
        "description": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "A description of what the response format is for, used by the model to determine how to respond in the format.",
          "title": "Description"
        },
        "schema": {
          "anyOf": [
            {
              "additionalProperties": true,
              "type": "object"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The schema for the response format, described as a JSON Schema object.",
          "title": "Schema"
        },
        "strict": {
          "anyOf": [
            {
              "type": "boolean"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Whether to enable strict schema adherence when generating the output.",
          "title": "Strict"
        }
      },
      "required": [
        "name"
      ],
      "title": "JsonSchema",
      "type": "object"
    },
    "MetaData": {
      "additionalProperties": false,
      "description": "Metadata associated with a value for Sequrity Control's policy enforcement system.\n\nEvery value flowing through the dual-LLM interpreter carries metadata that\nrecords *who* produced it, *who* may consume it, and arbitrary *tags* for\npolicy matching.",
      "properties": {
        "producers": {
          "items": {
            "type": "string"
          },
          "title": "Producers",
          "type": "array",
          "uniqueItems": true
        },
        "consumers": {
          "items": {
            "type": "string"
          },
          "title": "Consumers",
          "type": "array",
          "uniqueItems": true
        },
        "tags": {
          "items": {
            "type": "string"
          },
          "title": "Tags",
          "type": "array",
          "uniqueItems": true
        }
      },
      "title": "MetaData",
      "type": "object"
    },
    "ResponseFormatJsonObject": {
      "description": "JSON object response format.",
      "properties": {
        "type": {
          "const": "json_object",
          "description": "The type of response format being defined. Always 'json_object'.",
          "title": "Type",
          "type": "string"
        }
      },
      "required": [
        "type"
      ],
      "title": "ResponseFormatJsonObject",
      "type": "object"
    },
    "ResponseFormatJsonSchema": {
      "description": "JSON schema response format.",
      "properties": {
        "type": {
          "const": "json_schema",
          "description": "The type of response format being defined. Always 'json_schema'.",
          "title": "Type",
          "type": "string"
        },
        "json_schema": {
          "$ref": "#/$defs/JsonSchema",
          "description": "Structured Outputs configuration options, including a JSON Schema."
        }
      },
      "required": [
        "type",
        "json_schema"
      ],
      "title": "ResponseFormatJsonSchema",
      "type": "object"
    },
    "ResponseFormatText": {
      "description": "Text response format.",
      "properties": {
        "type": {
          "const": "text",
          "description": "The type of response format being defined. Always 'text'.",
          "title": "Type",
          "type": "string"
        }
      },
      "required": [
        "type"
      ],
      "title": "ResponseFormatText",
      "type": "object"
    },
    "SystemMessage": {
      "description": "System message.",
      "properties": {
        "role": {
          "const": "system",
          "description": "The role of the messages author, in this case 'system'.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "items": {
                "$ref": "#/$defs/ContentPartText"
              },
              "type": "array"
            }
          ],
          "description": "The contents of the system message.",
          "title": "Content"
        },
        "name": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "An optional name for the participant. Provides the model information to differentiate between participants of the same role.",
          "title": "Name"
        }
      },
      "required": [
        "role",
        "content"
      ],
      "title": "SystemMessage",
      "type": "object"
    },
    "ToolCall": {
      "description": "Tool call in assistant message.",
      "properties": {
        "id": {
          "description": "The ID of the tool call.",
          "title": "Id",
          "type": "string"
        },
        "type": {
          "const": "function",
          "description": "The type of the tool. Currently, only 'function' is supported.",
          "title": "Type",
          "type": "string"
        },
        "function": {
          "$ref": "#/$defs/ToolCallFunction",
          "description": "The function that the model called."
        }
      },
      "required": [
        "id",
        "type",
        "function"
      ],
      "title": "ToolCall",
      "type": "object"
    },
    "ToolCallFunction": {
      "description": "Function details in tool call.",
      "properties": {
        "arguments": {
          "description": "The arguments to call the function with, as generated by the model in JSON format.",
          "title": "Arguments",
          "type": "string"
        },
        "name": {
          "description": "The name of the function to call.",
          "title": "Name",
          "type": "string"
        }
      },
      "required": [
        "arguments",
        "name"
      ],
      "title": "ToolCallFunction",
      "type": "object"
    },
    "ToolMessage": {
      "description": "Tool message.",
      "properties": {
        "role": {
          "const": "tool",
          "description": "The role of the messages author, in this case 'tool'.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "items": {
                "$ref": "#/$defs/ContentPartText"
              },
              "type": "array"
            }
          ],
          "description": "The contents of the tool message.",
          "title": "Content"
        },
        "tool_call_id": {
          "description": "Tool call that this message is responding to.",
          "title": "Tool Call Id",
          "type": "string"
        }
      },
      "required": [
        "role",
        "content",
        "tool_call_id"
      ],
      "title": "ToolMessage",
      "type": "object"
    },
    "UserMessage": {
      "description": "User message.",
      "properties": {
        "role": {
          "const": "user",
          "description": "The role of the messages author, in this case 'user'.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "items": {
                "anyOf": [
                  {
                    "$ref": "#/$defs/ContentPartText"
                  },
                  {
                    "$ref": "#/$defs/ContentPartImage"
                  },
                  {
                    "$ref": "#/$defs/ContentPartInputAudio"
                  },
                  {
                    "$ref": "#/$defs/ContentPartFile"
                  }
                ]
              },
              "type": "array"
            }
          ],
          "description": "The contents of the user message.",
          "title": "Content"
        },
        "name": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "An optional name for the participant. Provides the model information to differentiate between participants of the same role.",
          "title": "Name"
        }
      },
      "required": [
        "role",
        "content"
      ],
      "title": "UserMessage",
      "type": "object"
    },
    "ValueWithMeta": {
      "description": "Python value wrapper with metadata for Sequrity Control's policy enforcement system.",
      "properties": {
        "value": {
          "title": "Value"
        },
        "meta": {
          "$ref": "#/$defs/MetaData"
        }
      },
      "required": [
        "value"
      ],
      "title": "ValueWithMeta",
      "type": "object"
    }
  },
  "description": "Extended chat completion request for LangGraph execution (OpenAI-compatible providers).",
  "properties": {
    "user_provided_program": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "The user-provided Python program to execute.",
      "title": "User Provided Program"
    },
    "context_vars": {
      "anyOf": [
        {
          "additionalProperties": {
            "$ref": "#/$defs/ValueWithMeta"
          },
          "type": "object"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "Predefined variables with meta for the interpreter.",
      "title": "Context Vars"
    },
    "messages": {
      "description": "A list of messages comprising the conversation so far.",
      "items": {
        "discriminator": {
          "mapping": {
            "assistant": "#/$defs/AssistantMessage",
            "developer": "#/$defs/DeveloperMessage",
            "function": "#/$defs/FunctionMessage",
            "system": "#/$defs/SystemMessage",
            "tool": "#/$defs/ToolMessage",
            "user": "#/$defs/UserMessage"
          },
          "propertyName": "role"
        },
        "oneOf": [
          {
            "$ref": "#/$defs/DeveloperMessage"
          },
          {
            "$ref": "#/$defs/SystemMessage"
          },
          {
            "$ref": "#/$defs/UserMessage"
          },
          {
            "$ref": "#/$defs/AssistantMessage"
          },
          {
            "$ref": "#/$defs/ToolMessage"
          },
          {
            "$ref": "#/$defs/FunctionMessage"
          }
        ]
      },
      "title": "Messages",
      "type": "array"
    },
    "model": {
      "description": "Model ID used to generate the response, like 'gpt-4o' or 'o3'.",
      "title": "Model",
      "type": "string"
    },
    "reasoning_effort": {
      "anyOf": [
        {
          "enum": [
            "none",
            "minimal",
            "low",
            "medium",
            "high",
            "xhigh"
          ],
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "Constrains effort on reasoning for reasoning models. Supported values are 'none', 'minimal', 'low', 'medium', 'high', and 'xhigh'.",
      "title": "Reasoning Effort"
    },
    "response_format": {
      "anyOf": [
        {
          "$ref": "#/$defs/ResponseFormatText"
        },
        {
          "$ref": "#/$defs/ResponseFormatJsonObject"
        },
        {
          "$ref": "#/$defs/ResponseFormatJsonSchema"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "An object specifying the format that the model must output. Setting to json_schema enables Structured Outputs.",
      "title": "Response Format"
    },
    "seed": {
      "anyOf": [
        {
          "type": "integer"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "If specified, the system will make a best effort to sample deterministically for reproducible results.",
      "title": "Seed"
    },
    "stream": {
      "anyOf": [
        {
          "type": "boolean"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "If set to true, partial message deltas will be sent as server-sent events.",
      "title": "Stream"
    },
    "temperature": {
      "anyOf": [
        {
          "type": "number"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "What sampling temperature to use, between 0 and 2. Higher values make output more random.",
      "title": "Temperature"
    },
    "tools": {
      "anyOf": [
        {
          "items": {
            "$ref": "#/$defs/FunctionTool"
          },
          "type": "array"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "A list of tools the model may call. Currently, only functions are supported as a tool.",
      "title": "Tools"
    },
    "top_p": {
      "anyOf": [
        {
          "type": "number"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "An alternative to sampling with temperature, called nucleus sampling. Use 0.1 for top 10% probability mass.",
      "title": "Top P"
    }
  },
  "required": [
    "messages",
    "model"
  ],
  "title": "LangGraphChatCompletionRequest",
  "type": "object"
}

Config:

  • extra: ignore

Fields:

user_provided_program pydantic-field 

user_provided_program: str | None = None

The user-provided Python program to execute.

context_vars pydantic-field 

context_vars: dict[str, ValueWithMeta] | None = None

Predefined variables with meta for the interpreter.

messages pydantic-field 

messages: list[Message]

A list of messages comprising the conversation so far.

model pydantic-field 

model: str

Model ID used to generate the response, like 'gpt-4o' or 'o3'.

reasoning_effort pydantic-field 

reasoning_effort: ReasoningEffort | None = None

Constrains effort on reasoning for reasoning models. Supported values are 'none', 'minimal', 'low', 'medium', 'high', and 'xhigh'.

response_format pydantic-field 

response_format: ResponseFormat | None = None

An object specifying the format that the model must output. Setting to json_schema enables Structured Outputs.

seed pydantic-field 

seed: int | None = None

If specified, the system will make a best effort to sample deterministically for reproducible results.

stream pydantic-field 

stream: bool | None = None

If set to true, partial message deltas will be sent as server-sent events.

temperature pydantic-field 

temperature: float | None = None

What sampling temperature to use, between 0 and 2. Higher values make output more random.

tools pydantic-field 

tools: list[Tool] | None = None

A list of tools the model may call. Currently, only functions are supported as a tool.

top_p pydantic-field 

top_p: float | None = None

An alternative to sampling with temperature, called nucleus sampling. Use 0.1 for top 10% probability mass.

LangGraphChatCompletionResponse pydantic-model

Chat completion response from LangGraph execution.

Show JSON schema: 
{
  "$defs": {
    "Annotation": {
      "description": "Annotation in message.",
      "properties": {
        "type": {
          "const": "url_citation",
          "description": "The type of the URL citation. Always 'url_citation'.",
          "title": "Type",
          "type": "string"
        },
        "url_citation": {
          "$ref": "#/$defs/AnnotationUrlCitation",
          "description": "A URL citation when using web search."
        }
      },
      "required": [
        "type",
        "url_citation"
      ],
      "title": "Annotation",
      "type": "object"
    },
    "AnnotationUrlCitation": {
      "description": "URL citation annotation.",
      "properties": {
        "end_index": {
          "description": "The index of the last character of the URL citation in the message.",
          "title": "End Index",
          "type": "integer"
        },
        "start_index": {
          "description": "The index of the first character of the URL citation in the message.",
          "title": "Start Index",
          "type": "integer"
        },
        "title": {
          "description": "The title of the web resource.",
          "title": "Title",
          "type": "string"
        },
        "url": {
          "description": "The URL of the web resource.",
          "title": "Url",
          "type": "string"
        }
      },
      "required": [
        "end_index",
        "start_index",
        "title",
        "url"
      ],
      "title": "AnnotationUrlCitation",
      "type": "object"
    },
    "AudioResponse": {
      "description": "Audio response from the model.",
      "properties": {
        "id": {
          "description": "Unique identifier for this audio response.",
          "title": "Id",
          "type": "string"
        },
        "data": {
          "description": "Base64 encoded audio bytes generated by the model, in the format specified in the request.",
          "title": "Data",
          "type": "string"
        },
        "expires_at": {
          "description": "The Unix timestamp (in seconds) for when this audio response will no longer be accessible on the server.",
          "title": "Expires At",
          "type": "integer"
        },
        "transcript": {
          "description": "Transcript of the audio generated by the model.",
          "title": "Transcript",
          "type": "string"
        }
      },
      "required": [
        "id",
        "data",
        "expires_at",
        "transcript"
      ],
      "title": "AudioResponse",
      "type": "object"
    },
    "Choice": {
      "description": "A chat completion choice.",
      "properties": {
        "finish_reason": {
          "anyOf": [
            {
              "enum": [
                "stop",
                "length",
                "tool_calls",
                "content_filter",
                "function_call",
                "error"
              ],
              "type": "string"
            },
            {
              "type": "string"
            }
          ],
          "description": "The reason the model stopped generating tokens. 'stop' for natural stop, 'length' for max tokens, 'tool_calls' for tool call, 'content_filter' for content filter, 'function_call' (deprecated) for function call, 'error' if an error occurred.",
          "title": "Finish Reason"
        },
        "index": {
          "description": "The index of the choice in the list of choices.",
          "title": "Index",
          "type": "integer"
        },
        "message": {
          "$ref": "#/$defs/ResponseMessage"
        },
        "logprobs": {
          "anyOf": [
            {
              "$ref": "#/$defs/ChoiceLogprobs"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Log probability information for the choice."
        }
      },
      "required": [
        "finish_reason",
        "index",
        "message"
      ],
      "title": "Choice",
      "type": "object"
    },
    "ChoiceLogprobs": {
      "description": "Log probability information for a choice.",
      "properties": {
        "content": {
          "anyOf": [
            {
              "items": {
                "$ref": "#/$defs/TokenLogprob"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "A list of message content tokens with log probability information.",
          "title": "Content"
        },
        "refusal": {
          "anyOf": [
            {
              "items": {
                "$ref": "#/$defs/TokenLogprob"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "A list of message refusal tokens with log probability information.",
          "title": "Refusal"
        }
      },
      "title": "ChoiceLogprobs",
      "type": "object"
    },
    "CompletionUsage": {
      "description": "Usage statistics for the completion request.",
      "properties": {
        "completion_tokens": {
          "description": "Number of tokens in the generated completion.",
          "title": "Completion Tokens",
          "type": "integer"
        },
        "prompt_tokens": {
          "description": "Number of tokens in the prompt.",
          "title": "Prompt Tokens",
          "type": "integer"
        },
        "total_tokens": {
          "description": "Total number of tokens used in the request (prompt + completion).",
          "title": "Total Tokens",
          "type": "integer"
        }
      },
      "required": [
        "completion_tokens",
        "prompt_tokens",
        "total_tokens"
      ],
      "title": "CompletionUsage",
      "type": "object"
    },
    "ResponseMessage": {
      "description": "A chat completion message generated by the model.",
      "properties": {
        "role": {
          "const": "assistant",
          "description": "The role of the author of this message.",
          "title": "Role",
          "type": "string"
        },
        "content": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The contents of the message.",
          "title": "Content"
        },
        "refusal": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The refusal message generated by the model.",
          "title": "Refusal"
        },
        "annotations": {
          "anyOf": [
            {
              "items": {
                "$ref": "#/$defs/Annotation"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "Annotations for the message, when applicable, as when using the web search tool.",
          "title": "Annotations"
        },
        "audio": {
          "anyOf": [
            {
              "$ref": "#/$defs/AudioResponse"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "If the audio output modality is requested, this object contains data about the audio response from the model."
        },
        "tool_calls": {
          "anyOf": [
            {
              "items": {
                "$ref": "#/$defs/ResponseToolCall"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "The tool calls generated by the model, such as function calls.",
          "title": "Tool Calls"
        }
      },
      "required": [
        "role"
      ],
      "title": "ResponseMessage",
      "type": "object"
    },
    "ResponseToolCall": {
      "description": "Tool call in assistant message.",
      "properties": {
        "id": {
          "description": "The ID of the tool call.",
          "title": "Id",
          "type": "string"
        },
        "type": {
          "const": "function",
          "description": "The type of the tool. Currently, only 'function' is supported.",
          "title": "Type",
          "type": "string"
        },
        "function": {
          "$ref": "#/$defs/ToolCallFunction",
          "description": "The function that the model called."
        }
      },
      "required": [
        "id",
        "type",
        "function"
      ],
      "title": "ResponseToolCall",
      "type": "object"
    },
    "TokenLogprob": {
      "description": "Token log probability information.",
      "properties": {
        "token": {
          "description": "The token.",
          "title": "Token",
          "type": "string"
        },
        "bytes": {
          "anyOf": [
            {
              "items": {
                "type": "integer"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "A list of integers representing the UTF-8 bytes representation of the token.",
          "title": "Bytes"
        },
        "logprob": {
          "description": "The log probability of this token, if it is within the top 20 most likely tokens.",
          "title": "Logprob",
          "type": "number"
        },
        "top_logprobs": {
          "description": "List of the most likely tokens and their log probability, at this token position.",
          "items": {
            "$ref": "#/$defs/TopLogprob"
          },
          "title": "Top Logprobs",
          "type": "array"
        }
      },
      "required": [
        "token",
        "logprob",
        "top_logprobs"
      ],
      "title": "TokenLogprob",
      "type": "object"
    },
    "ToolCallFunction": {
      "description": "Function details in tool call.",
      "properties": {
        "arguments": {
          "description": "The arguments to call the function with, as generated by the model in JSON format.",
          "title": "Arguments",
          "type": "string"
        },
        "name": {
          "description": "The name of the function to call.",
          "title": "Name",
          "type": "string"
        }
      },
      "required": [
        "arguments",
        "name"
      ],
      "title": "ToolCallFunction",
      "type": "object"
    },
    "TopLogprob": {
      "description": "Top logprob entry.",
      "properties": {
        "token": {
          "description": "The token.",
          "title": "Token",
          "type": "string"
        },
        "bytes": {
          "anyOf": [
            {
              "items": {
                "type": "integer"
              },
              "type": "array"
            },
            {
              "type": "null"
            }
          ],
          "default": null,
          "description": "A list of integers representing the UTF-8 bytes representation of the token.",
          "title": "Bytes"
        },
        "logprob": {
          "description": "The log probability of this token, if it is within the top 20 most likely tokens.",
          "title": "Logprob",
          "type": "number"
        }
      },
      "required": [
        "token",
        "logprob"
      ],
      "title": "TopLogprob",
      "type": "object"
    }
  },
  "description": "Chat completion response from LangGraph execution.",
  "properties": {
    "id": {
      "description": "A unique identifier for the chat completion.",
      "title": "Id",
      "type": "string"
    },
    "choices": {
      "description": "A list of chat completion choices. Can be more than one if 'n' is greater than 1.",
      "items": {
        "$ref": "#/$defs/Choice"
      },
      "title": "Choices",
      "type": "array"
    },
    "created": {
      "description": "The Unix timestamp (in seconds) of when the chat completion was created.",
      "title": "Created",
      "type": "integer"
    },
    "model": {
      "description": "The model used for the chat completion.",
      "title": "Model",
      "type": "string"
    },
    "object": {
      "const": "chat.completion",
      "description": "The object type, which is always 'chat.completion'.",
      "title": "Object",
      "type": "string"
    },
    "usage": {
      "anyOf": [
        {
          "$ref": "#/$defs/CompletionUsage"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "Usage statistics for the completion request."
    },
    "session_id": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "default": null,
      "description": "The session ID associated with this chat completion, if applicable.",
      "title": "Session Id"
    }
  },
  "required": [
    "id",
    "choices",
    "created",
    "model",
    "object"
  ],
  "title": "LangGraphChatCompletionResponse",
  "type": "object"
}

Config:

  • extra: ignore

Fields:

id pydantic-field 

id: str

A unique identifier for the chat completion.

choices pydantic-field 

choices: list[Choice]

A list of chat completion choices. Can be more than one if 'n' is greater than 1.

created pydantic-field 

created: int

The Unix timestamp (in seconds) of when the chat completion was created.

model pydantic-field 

model: str

The model used for the chat completion.

object pydantic-field 

object: Literal['chat.completion']

The object type, which is always 'chat.completion'.

usage pydantic-field 

usage: CompletionUsage | None = None

Usage statistics for the completion request.

session_id pydantic-field 

session_id: str | None = None

The session ID associated with this chat completion, if applicable.

LangGraphExecutor 

LangGraphExecutor(
    graph: StateGraph,
    node_functions: dict[str, Callable] | None = None,
    internal_node_mapping: dict[str, str] | None = None,
)

Executor for running LangGraph StateGraphs through Sequrity.

This class handles
  1. Converting LangGraph to executable Python code
  2. Mapping nodes to tools (external) or internal functions
  3. Executing code via secure orchestrator
  4. Handling tool calls for external nodes

Parameters:

  • graph

    (StateGraph) –

    LangGraph StateGraph to execute.

  • node_functions

    (dict[str, Callable] | None, default: None ) –

    Dict mapping node names to their functions.

  • internal_node_mapping

    (dict[str, str] | None, default: None ) –

    Map of node names to internal tool names (e.g., {"agent": "parse_with_ai"}).

Raises:

Methods:

build_tool_definitions 

build_tool_definitions(rest_api_type: RestApiType = CHAT_COMPLETIONS) -> list[dict]

Build tool definitions for external nodes.

Parameters:

  • rest_api_type

    (RestApiType, default: CHAT_COMPLETIONS ) –

    The REST API type determining the tool definition format. CHAT_COMPLETIONS produces OpenAI-style definitions. MESSAGES produces Anthropic-style definitions.

Returns:

  • list[dict] –

    List of tool definition dicts in the appropriate format.

execute_tool_call 

execute_tool_call(
    tool_call: dict, rest_api_type: RestApiType = CHAT_COMPLETIONS
) -> dict

Execute a tool call (external node).

Parameters:

  • tool_call

    (dict) –

    Tool call dict. Format depends on rest_api_type: - CHAT_COMPLETIONS: {"function": {"name": ..., "arguments": ...}} - MESSAGES: {"name": ..., "input": {...}}

  • rest_api_type

    (RestApiType, default: CHAT_COMPLETIONS ) –

    The REST API type determining the tool call format.

Returns:

  • dict –

    The result dict from executing the node function.

Raises:

  • RuntimeError –

    If no function is found for the specified tool name.