API Documentation: Create Chat Function Calling

Interface Description

Given a prompt, the model returns one or more predicted completions and can also return the probabilities of alternative tokens at each position. This interface specifically supports the Function Calling feature, allowing the model to choose intelligently between generating a message and calling a specified function.

• Official Reference: Function Calling Guide

Base Information

• Request Method: POST
• Request URL: https://api.codingplanx.ai/v1/chat/completions
• Data Format: application/json

---

Request Headers

---

Request Body

Request Body Example

{
    "model": "gpt-4o",
    "messages": [
        {
            "role": "system",
            "content": "You are a helpful customer support assistant. Use the supplied tools to assist the user."
        },
        {
            "role": "user",
            "content": "Hi, can you tell me the delivery date for my order?"
        },
        {"role": "assistant", "content": "Hi there! I can help with that. Can you please provide your order ID?"},
        {"role": "user", "content": "i think it is order_12345"}
    ],
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "get_delivery_date",
                "description": "Get the delivery date for a customer's order. Call this whenever you need to know the delivery date, for example when a customer asks 'Where is my package'",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "order_id": {
                            "type": "string",
                            "description": "The customer's order ID."
                        }
                    },
                    "required": [
                        "order_id"
                    ],
                    "additionalProperties": false
                }
            }
        }
    ]
}

---

Response Parameters

Response Example (200 OK)

{
    "id": "chatcmpl-123",
    "object": "chat.completion",
    "created": 1677652288,
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "\r
\r
Hello there, how may I assist you today?"
            },
            "finish_reason": "stop"
        }
    ],
    "usage": {
        "prompt_tokens": 9,
        "completion_tokens": 12,
        "total_tokens": 21
    }
}

---

FAQs

Q1: How do I force the model to call a specific function? A: You can achieve this by setting the tool_choice parameter. For example, if your function name is get_weather, set tool_choice to: {"type": "function", "function": {"name": "get_weather"}}. This forces the model to ignore a standard reply and output the arguments for that specific function.

Q2: Why is my response content truncated? A: Check the value of choices[0].finish_reason in the response. If it is "length", it means the generated text exceeded the max_tokens limit or the model's maximum context length. You can try increasing the max_tokens value.

Q3: Both temperature and top_p control randomness; how should I use them? A: Both parameters adjust the diversity of the output. Higher values make the output more creative, while lower values make it more deterministic. Official documentation recommends adjusting only one of these parameters and keeping the other at its default value to avoid unpredictable results.

Q4: What is the data format when stream: true is enabled? A: When streaming is enabled, the interface no longer returns a single JSON object. Instead, it returns a stream of Server-Sent Events (SSE). Data is sent in chunks as data: {"id": "...", "choices": [{"delta": {"content": "..."}}]}. The final message will be data: [DONE].

Q5: What should I keep in mind when enabling response_format for JSON mode? A: When setting response_format to { "type": "json_object" }, you must also explicitly instruct the model via a message (system or user prompt) to generate JSON. Failure to do so may cause the model to output endless whitespace, leading to request timeouts or token limit exhaustion.