Chat Completions

Endpoint address

POST https://api.yomiapi.com/v1/chat/completions

Request headers:

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Request parameters

The endpoint also supports the remaining standard OpenAI Chat Completions parameters (such as stop, frequency_penalty). Include them as needed.

messages format

Each message contains role and content.system defines the AI's role and behavior, user is user input, and assistant contains previous model replies:

[
  {"role": "system", "content": "You are a helpful assistant"},
  {"role": "user", "content": "Hello"},
  {"role": "assistant", "content": "Hello! How can I help you?"},
  {"role": "user", "content": "Please introduce yourself"}
]

Code examples

Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.yomiapi.com/v1"
)
response = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[
        {"role": "system", "content": "You are a helpful assistant"},
        {"role": "user", "content": "Hello, please introduce yourself"}
    ],
    temperature=0.7,
    max_tokens=1000
)
print(response.choices[0].message.content)

Node.js

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'YOUR_API_KEY',
  baseURL: 'https://api.yomiapi.com/v1'
});

async function main() {
  const response = await client.chat.completions.create({
    model: 'deepseek-v4-flash',
    messages: [
      { role: 'system', content: 'You are a helpful assistant' },
      { role: 'user', content: 'Hello, please introduce yourself' }
    ],
    temperature: 0.7,
    max_tokens: 1000
  });
  console.log(response.choices[0].message.content);
}
main();

cURL

curl https://api.yomiapi.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "deepseek-v4-flash",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant"},
      {"role": "user", "content": "Hello, please introduce yourself"}
    ],
    "temperature": 0.7,
    "max_tokens": 1000
  }'

Response format

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1234567890,
  "model": "deepseek-v4-flash",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "Hello! I am an AI assistant..."
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 25,
    "completion_tokens": 100,
    "total_tokens": 125
  }
}

The reply is in choices[0].message.content;finish_reason is stop indicates normal completion, while length indicates truncation at the max_tokens limit.

Streaming output

Set stream: true to receive incremental responses in Server-Sent Events (SSE) format:

data: {"id":"chatcmpl-123","object":"chat.completion.chunk","choices":[{"delta":{"content":"Hello"}}]}
data: {"id":"chatcmpl-123","object":"chat.completion.chunk","choices":[{"delta":{"content":" world"}}]}
data: [DONE]

When using the SDK, read each increment:

response = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Write a poem"}],
    stream=True
)
for chunk in response:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

Multimodal input

Multimodal models (such as claude-sonnet-5, gpt-5.6-sol) can receive a mix of text, images and other content parts in content, using the same /v1/chat/completions endpoint described here. See Image understanding (vision) and Video understanding for the exact format.

Recommendations

• System prompt: Use the system role to set the AI's identity, tone and output format. This is more stable than repeating the requirements in every user message.
• Context management: Include historical messages in multi-turn conversations. Trim or summarize older content when the conversation grows long to save tokens.
• Streaming: For long text generated for end users, enable stream to substantially reduce time to first token.
• Error handling: See API overview and authentication for error codes and retry strategies.