Anthropic Core Workflow A — Tool Use (Function Calling)

Overview

Implement Claude's tool use capability where the model can call functions you define. Claude returns tool_use content blocks with structured JSON inputs; your code executes the function and returns tool_result blocks. This is the foundation for building AI agents.

Prerequisites

• Completed anth-install-auth setup
• Understanding of the Messages API request/response cycle
• Functions or APIs you want Claude to call

Instructions

Step 1: Define Tools

import anthropic

client = anthropic.Anthropic()

tools = [
    {
        "name": "get_weather",
        "description": "Get current weather for a city. Use when the user asks about weather conditions.",
        "input_schema": {
            "type": "object",
            "properties": {
                "city": {
                    "type": "string",
                    "description": "City name, e.g. 'San Francisco, CA'"
                },
                "units": {
                    "type": "string",
                    "enum": ["celsius", "fahrenheit"],
                    "description": "Temperature units"
                }
            },
            "required": ["city"]
        }
    },
    {
        "name": "search_database",
        "description": "Search product database by query string. Returns matching products.",
        "input_schema": {
            "type": "object",
            "properties": {
                "query": {"type": "string"},
                "max_results": {"type": "integer", "default": 10}
            },
            "required": ["query"]
        }
    }
]

Step 2: Send Request with Tools

message = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    tools=tools,
    messages=[{"role": "user", "content": "What's the weather in Tokyo?"}]
)

# Claude responds with stop_reason="tool_use"
# message.content contains both text and tool_use blocks:
# [
#   {"type": "text", "text": "I'll check the weather for you."},
#   {"type": "tool_use", "id": "toolu_01A...", "name": "get_weather",
#    "input": {"city": "Tokyo", "units": "celsius"}}
# ]

Step 3: Execute Tool and Return Result

def execute_tool(name: str, input_data: dict) -> str:
    """Route tool calls to actual implementations."""
    if name == "get_weather":
        # Call your weather API
        return '{"temp": 22, "condition": "partly cloudy", "humidity": 65}'
    elif name == "search_database":
        return '{"results": [{"name": "Widget A", "price": 29.99}]}'
    raise ValueError(f"Unknown tool: {name}")

# Extract tool_use blocks and execute
tool_results = []
for block in message.content:
    if block.type == "tool_use":
        result = execute_tool(block.name, block.input)
        tool_results.append({
            "type": "tool_result",
            "tool_use_id": block.id,  # Must match the tool_use block id
            "content": result
        })

# Continue conversation with tool results
follow_up = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    tools=tools,
    messages=[
        {"role": "user", "content": "What's the weather in Tokyo?"},
        {"role": "assistant", "content": message.content},
        {"role": "user", "content": tool_results}
    ]
)

print(follow_up.content[0].text)
# "The current weather in Tokyo is 22°C and partly cloudy with 65% humidity."

Step 4: Agentic Loop (Multiple Tool Calls)

def run_agent(user_message: str, tools: list, max_turns: int = 10) -> str:
    """Run an agentic loop that handles multiple sequential tool calls."""
    messages = [{"role": "user", "content": user_message}]

    for _ in range(max_turns):
        response = client.messages.create(
            model="claude-sonnet-4-20250514",
            max_tokens=4096,
            tools=tools,
            messages=messages
        )

        # If Claude is done (no more tool calls), return final text
        if response.stop_reason == "end_turn":
            return next(
                (b.text for b in response.content if b.type == "text"), ""
            )

        # Process tool calls
        messages.append({"role": "assistant", "content": response.content})
        tool_results = []
        for block in response.content:
            if block.type == "tool_use":
                result = execute_tool(block.name, block.input)
                tool_results.append({
                    "type": "tool_result",
                    "tool_use_id": block.id,
                    "content": result
                })
        messages.append({"role": "user", "content": tool_results})

    return "Max turns reached"

Output

• Tool definitions with JSON Schema input validation
• Agent loop handling sequential tool calls
• Proper tool_use / tool_result message threading

Error Handling

Tool Choice Options

# Let Claude decide (default)
tool_choice={"type": "auto"}

# Force Claude to use a specific tool
tool_choice={"type": "tool", "name": "get_weather"}

# Force Claude to use any tool (must call at least one)
tool_choice={"type": "any"}

Resources

• Tool Use Guide
• Tool Use API Reference
• Tool Use Examples

Next Steps

For streaming with tools, see anth-core-workflow-b.