Implementation:Cohere ai Cohere python ToolCall Model

Knowledge Sources	Cohere Python SDK
Domains	SDK, Function Calling
Last Updated	2026-02-15 14:00 GMT

Overview

ToolCall is a Pydantic model representing a model-generated tool call request, containing the tool name and the parameters to invoke it with.

Description

The ToolCall model encapsulates a function call generated by the Cohere chat model during the function calling workflow. When tools are provided to the Chat API and the model determines that a tool should be invoked, it returns one or more ToolCall objects.

Each ToolCall contains:

• name -- The name of the tool to invoke, matching one of the tool definitions provided in the request.
• parameters -- A dictionary of parameter name-value pairs to pass to the tool function.

The model extends UncheckedBaseModel and allows extra fields.

Usage

Use ToolCall when processing chat responses that include function calling. After receiving tool calls from the model, use the name to dispatch to the appropriate function and pass the parameters as arguments. Then submit the tool results back to the model to continue the conversation.

Code Reference

Source Location

• Repository: Cohere Python SDK
• File: src/cohere/types/tool_call.py

Signature

class ToolCall(UncheckedBaseModel):
    name: str = pydantic.Field()
    parameters: typing.Dict[str, typing.Any] = pydantic.Field()

Import

from cohere.types import ToolCall

I/O Contract

Fields

Field	Type	Required	Description
name	str	Yes	Name of the tool to call. Matches a tool name from the tool definitions provided in the request.
parameters	Dict[str, Any]	Yes	The name-value pairs of parameters to use when invoking the tool.

Usage Examples

from cohere.types import ToolCall, Tool

# Define tools and make a chat request
tools = [
    Tool(
        name="search_database",
        description="Search a product database by query.",
        parameter_definitions={
            "query": {"description": "Search query", "type": "str", "required": True}
        }
    )
]

response = client.chat(
    message="Find me running shoes",
    tools=tools
)

# Process tool calls from the response
if response.tool_calls:
    for tool_call in response.tool_calls:
        print(f"Tool: {tool_call.name}")
        print(f"Parameters: {tool_call.parameters}")

        # Dispatch to the appropriate function
        if tool_call.name == "search_database":
            result = search_database(**tool_call.parameters)

            # Submit tool results back to the model
            response = client.chat(
                message="Find me running shoes",
                tools=tools,
                tool_results=[{
                    "call": tool_call,
                    "outputs": [result]
                }]
            )

Related Pages

• Environment:Cohere_ai_Cohere_python_Python_SDK_Runtime