# OpenAI Chat Completions

OpenAI Chat Completions

## Endpoint and authentication

`POST https://api.tu-zi.com/v1/chat/completions`

Transport: json

Use an inference API key. Examples contain placeholders; read credentials from environment variables.

```http
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
```

## Body parameters

| Field | Type | Required | Description / constraints |
| --- | --- | --- | --- |
| model | string | Yes | Model identifier; use the currently selected model. Example: gpt-4o-mini |
| messages | array | Yes | Conversation messages, including at least one user message. Example: [{"role":"user","content":"你好"}] |
| stream | boolean | No | Enable incremental response delivery. Disable to receive the complete result. This controls transport rather than generated content. Form initial value (not an upstream default): false Example: false |
| temperature | number | No | Sampling randomness: lower values are more stable, higher values more varied. Usually adjust either temperature or top_p. Form initial value (not an upstream default): 0.7 Range in this definition: 0 ~ 2 Example: 0.7 |
| max_tokens | number | No | Maximum output token budget; reasoning may consume part of the budget. Too small a value can truncate output. A larger limit can increase maximum time and cost. Form initial value (not an upstream default): 512 Range in this definition: 1 ~ 128000 Example: 512 |
| top_p | number | No | Cumulative token probability threshold. Lower values narrow the candidate set. Usually adjust either top_p or temperature. Form initial value (not an upstream default): 1 Range in this definition: 0 ~ 1 Example: 1 |
| frequency_penalty | number | No | Penalize repeated tokens based on frequency. Positive values discourage repetition; excessive values can harm coherence. Form initial value (not an upstream default): 0 Range in this definition: -2 ~ 2 Example: 0 |
| presence_penalty | number | No | Penalize tokens already present. Positive values encourage new content; excessive values can cause topic drift. Form initial value (not an upstream default): 0 Range in this definition: -2 ~ 2 Example: 0 |
| messages[].role | string | No | Message role: system, developer, user, assistant or tool; support depends on the model. |
| messages[].content | string \| array \| null | No | Text or multimodal blocks; image input uses image_url. Tool-call messages may have null content. |
| tools | array | No | Tool declarations when supported by the model; functions define name, description and parameters (JSON Schema). |
| tool_choice | string \| object | No | Tool-selection policy; must match tools and model capabilities. |
| response_format | object | No | Structured-output configuration; json_object/json_schema depend on model support. |
| stream_options | object | No | Streaming options; include_usage requires endpoint/model support. |

## Minimal request example

```json
{
  "model": "YOUR_MODEL",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ],
  "stream": false
}
```

## Before running

Replace https://api.tu-zi.com with this site origin (do not append /v1 twice), and YOUR_MODEL with the exact marketplace model ID. Install requests for Python; use Node.js 20+ and save JavaScript as .mjs. WebSocket samples list their own dependencies.

Create an inference key on the site and set API_KEY in the environment; replace YOUR_API_KEY in cURL. Prepare local input.png/input.wav files for upload examples.

Start with a minimal request, then read the result using the response fields below. For async APIs, a returned task ID requires polling; acceptance does not mean generation has finished.

## Code examples

### curl

```bash
curl --request POST 'https://api.tu-zi.com/v1/chat/completions' \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
  "model": "YOUR_MODEL",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ],
  "stream": false
}'
```

### python

```python
import json
import os
import requests

url = "https://api.tu-zi.com/v1/chat/completions"
headers = json.loads("{\"Authorization\":\"Bearer YOUR_API_KEY\",\"Content-Type\":\"application/json\"}")
headers["Authorization"] = "Bearer " + os.environ["API_KEY"]
payload = json.loads("{\"model\":\"YOUR_MODEL\",\"messages\":[{\"role\":\"user\",\"content\":\"Hello\"}],\"stream\":false}")
response = requests.request("POST", url, headers=headers, json=payload, timeout=120)
response.raise_for_status()
result = response.json()
if isinstance(result, dict) and (result.get("success") is False or result.get("error")):
    raise RuntimeError(json.dumps(result, ensure_ascii=False))
print(result)  # Also inspect protocol-specific status/code fields.
```

### javascript

```javascript
// Node.js 20+
const url = "https://api.tu-zi.com/v1/chat/completions";
const headers = {
  "Authorization": "Bearer YOUR_API_KEY",
  "Content-Type": "application/json"
};
headers.Authorization = 'Bearer ' + process.env.API_KEY;
const response = await fetch(url, { method: "POST", headers, body: JSON.stringify({
  "model": "YOUR_MODEL",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ],
  "stream": false
}) });
if (!response.ok) throw new Error(await response.text());
const result = await response.json();
if (result?.success === false || result?.error) throw new Error(JSON.stringify(result));
console.log(result); // Also inspect protocol-specific status/code fields.
```

## Response fields

| Field | Type | Description / constraints |
| --- | --- | --- |
| choices[].message.content | string \| array \| null | Generated content; may be null for tool calls. |
| choices[].finish_reason | string | Completion reason. |
| usage | object | Usage for this request. |
| id | string | Response ID. |
| model | string | Response model identifier. |
| choices[].message.role | string | Output message role. |
| choices[].message.tool_calls[] | object | Tool calls: inspect id, type, function.name and function.arguments. Arguments are a JSON string; validate before executing tools. |
| usage.prompt_tokens | integer | Input token count. |
| usage.completion_tokens | integer | Output token count. |
| usage.total_tokens | integer | Total token count. |

Illustrative response; some endpoints show essential fields only. Optional fields may be absent.

```json
{
  "id": "example",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Hello"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 5,
    "completion_tokens": 1,
    "total_tokens": 6
  }
}
```

## Streaming

With stream=true parse SSE choices[].delta until data: [DONE].

## Errors and retries

Check HTTP status and protocol-specific errors; do not mix error formats across protocols. For 401/403 check credentials and permissions, for 429 back off, and for 5xx check gateway/upstream status. Error bodies may contain error/message/code depending on the protocol.

Use bounded exponential backoff for reads. Do not blindly retry task creation, generation, or key mutations, to avoid duplicate billing or execution.

## Constraints

This page describes the base protocol. Use the selected model quick start for supported fields, enums, limits and asynchronous behavior. Replace YOUR_MODEL. Form initial values are not upstream defaults.

## Documentation source

Definition version: 1

Reviewed: 2026-11-02

Source: https://api.tu-zi.com/en/docs/api/reference-relay-openai

Related protocol guide: https://api.tu-zi.com/docs/api/openai

This export contains documentation and examples only, without real credentials, account data or user requests.