Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.llm7.io/llms.txt

Use this file to discover all available pages before exploring further.

Function calling is available on paid plans only. Add a token from token.llm7.io before you start.
Bind model outputs to your own logic with tool calls. This guide shows how to declare a function, let the model request it, and return the result to finish the conversation.

​
Prerequisites

  • Python 3.9+ with requests installed
  • A paid LLM7.io token set in API_KEY
  • Access to https://api.llm7.io/v1/chat/completions

​
Full example

This end-to-end example asks for the weather in London, lets the model request your get_weather tool, and sends back mocked data. 
import json
import requests

API_KEY = "your-token"  # Required for paid features
BASE_URL = "https://api.llm7.io/v1/chat/completions"

# 1) Describe the callable function
tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Get current weather for a location",
        "parameters": {
            "type": "object",
            "properties": {
                "location": {
                    "type": "string",
                    "description": "City and state, e.g. Boston, MA"
                },
                "unit": {
                    "type": "string",
                    "enum": ["celsius", "fahrenheit"]
                }
            },
            "required": ["location"]
        }
    }
}]

# 2) Let the model decide whether to call the function
payload = {
    "model": "openai",
    "messages": [{"role": "user", "content": "What's the weather in London?"}],
    "tools": tools,
    "tool_choice": "auto"
}

response = requests.post(
    BASE_URL,
    headers={"Authorization": f"Bearer {API_KEY}"},
    json=payload,
    timeout=30,
).json()

message = response["choices"][0]["message"]

# 3) If the model asks for a tool, supply the result and get the final reply
if message.get("tool_calls"):
    tool_call = message["tool_calls"][0]
    print(f"Model requested: {tool_call}")

    # Replace with a real weather API call
    weather_data = {
        "temperature": 12,
        "condition": "sunny",
        "unit": "celsius"
    }

    follow_up_messages = [
        {"role": "user", "content": "What's the weather in London?"},
        message,
        {
            "role": "tool",
            "tool_call_id": tool_call["id"],
            "content": json.dumps(weather_data)
        }
    ]

    final = requests.post(
        BASE_URL,
        headers={"Authorization": f"Bearer {API_KEY}"},
        json={"model": "openai", "messages": follow_up_messages},
        timeout=30,
    ).json()

    print(final["choices"][0]["message"]["content"])
else:
    print(message["content"])

​
Key parameters

  • tools: JSON Schema that declares each function you will allow the model to call.
  • tool_choice: "auto": Lets the model decide whether to call your tool or answer directly.
  • tool_calls[*].id: Connects the model’s request to your function output; return it in the tool_call_id field.
  • messages: Keep the full conversation history plus the tool call and its response so the model can produce a final answer.