By Ryze AI Team
This tutorial builds a minimal agent that can call a weather function to answer questions, using the OpenAI function calling API.
tools = [{
"type": "function",
"function": {
"name": "get_weather",
"description": "Get current weather for a city",
"parameters": {
"type": "object",
"properties": {"city": {"type": "string"}},
"required": ["city"]
}
}
}]
import json
def get_weather(city):
# call a weather API here
return {"city": city, "temp_c": 18, "condition": "cloudy"}
from openai import OpenAI
client = OpenAI()
messages = [{"role": "user", "content": "What is the weather in Berlin?"}]
for _ in range(10):
r = client.chat.completions.create(model="gpt-4o-mini", messages=messages, tools=tools)
msg = r.choices[0].message
messages.append(msg)
if not msg.tool_calls:
print(msg.content)
break
for tc in msg.tool_calls:
args = json.loads(tc.function.arguments)
result = get_weather(**args)
messages.append({"role": "tool", "tool_call_id": tc.id, "content": json.dumps(result)})
The model decides whether to call a tool. If it does, you execute it and feed the result back as a tool message. The loop continues until the model produces a normal reply. The cap of ten iterations prevents infinite loops.
Validate arguments before calling tools. Wrap tool execution in try/except and return an error message the model can read. Require confirmation for any mutating action. Log every call. For read-only tools, run automatically; for anything that changes state, ask the user first.
Docs: OpenAI function calling.