Tool Calling

Key idea

The model never executes anything. It only requests a call — your code decides whether to honour it, runs it, and reports the result back.

A chat model gains access to tools by binding them, then the round trip runs in your code, not inside the model:

from langchain.chat_models import init_chat_model

def get_weather(location: str) -> str:
    """Get the weather at a location."""
    return f"It's sunny in {location}."

model = init_chat_model("claude-sonnet-4-6", model_provider="anthropic")
model_with_tools = model.bind_tools([get_weather])

messages = [{"role": "user", "content": "What's the weather in Lisbon?"}]
ai_msg = model_with_tools.invoke(messages)
messages.append(ai_msg)

for tool_call in ai_msg.tool_calls:
    tool_result = get_weather.invoke(tool_call)
    messages.append(tool_result)

final_response = model_with_tools.invoke(messages)
print(final_response.text)

The AIMessage returned by the first invoke carries tool_calls instead of (or alongside) text — each entry has a name, args, and id. Your code executes the matching tool and appends a ToolMessage keyed to that id. The second invoke gives the model the result so it can produce a final answer.

Step	Who acts	What happens
1	model	emits `tool_calls` on an `AIMessage` instead of answering directly
2	your code	matches each call to a tool, executes it
3	your code	appends a `ToolMessage` per call, `tool_call_id` matching
4	model	reads the `ToolMessage`s, produces the final response

This manual loop is what create_agent automates: bind, call, execute, append, repeat until the model stops requesting tools. See Agent Concepts for when that automation is worth reaching for.

See also​

See also