Implementation:Mistralai Client python ToolMessage Model

Knowledge Sources	Mistral Client Python Mistral AI Function Calling
Domains	Function_Calling, API_Design
Last Updated	2026-02-15 14:00 GMT

Overview

Concrete tool for packaging function execution results into tool messages for submission back to the model provided by the ToolMessage model.

Description

The ToolMessage Pydantic model wraps function execution results for the multi-turn tool calling loop. It has a fixed role of "tool", a content field (string or content chunks) containing the function result, a tool_call_id that must match the ToolCall.id from the model's response, and an optional name field for the function name.

Usage

Create a ToolMessage for each executed tool call result. Include it in the messages list alongside the original conversation and the assistant's tool-calling message when making the follow-up chat.complete() call.

Code Reference

Source Location

• Repository: client-python
• File: src/mistralai/client/models/toolmessage.py
• Lines: L1-76

Signature

class ToolMessage(BaseModel):
    content: Union[str, List[ContentChunk]]
    tool_call_id: Optional[str] = None
    name: Optional[str] = None
    role: Literal["tool"] = "tool"

Import

from mistralai.models import ToolMessage

I/O Contract

Inputs

Name	Type	Required	Description
content	Union[str, List[ContentChunk]]	Yes	Function execution result as string
tool_call_id	Optional[str]	Yes (practical)	Must match ToolCall.id from model response
name	Optional[str]	No	Function name

Outputs

Name	Type	Description
ToolMessage object	ToolMessage	Validated tool result message for follow-up API call

Usage Examples

Complete Function Calling Loop

import json
from mistralai.models import UserMessage, ToolMessage

# Step 1: Initial request with tools
messages = [UserMessage(content="What's the status of transaction T001?")]
response = client.chat.complete(
    model="mistral-large-latest",
    messages=messages,
    tools=[payment_status_tool],
)

# Step 2: Parse tool calls and execute
assistant_msg = response.choices[0].message
tool_messages = []
for tool_call in assistant_msg.tool_calls:
    fn = registry[tool_call.function.name]
    args = json.loads(tool_call.function.arguments)
    result = fn(**args)
    tool_messages.append(
        ToolMessage(
            content=result,
            tool_call_id=tool_call.id,
            name=tool_call.function.name,
        )
    )

# Step 3: Follow-up request with tool results
messages = messages + [assistant_msg] + tool_messages
final_response = client.chat.complete(
    model="mistral-large-latest",
    messages=messages,
    tools=[payment_status_tool],
)

print(final_response.choices[0].message.content)

Related Pages

Implements Principle

• Principle:Mistralai_Client_python_Tool_Result_Submission