Skip to main content
APIYI is an OpenAI-compatible AI gateway: one standard interface and one API Key let you call 400+ mainstream large models. This page is a navigation hub — it helps you quickly find which model to use, test endpoints online, and learn how to integrate.

​
Platform Overview

​
OpenAI Compatible Mode

APIYI uses the OpenAI-compatible format. Once it works, switching models only means changing the model field — everything else stays the same: 
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.apiyi.com/v1"
)

# Switching models = change only the `model` field, nothing else
response = client.chat.completions.create(
    model="gpt-5-chat-latest",   # swap in any supported model name
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)
For exact model names, pricing, and recommended use cases, see the two dedicated pages under “Choose a Model” below. We don’t list them here to avoid stale information.

​
Feature Support Scope

Supported

  • Chat Completions
  • Image / video generation
  • Speech transcription (Whisper)
  • Embeddings
  • Function Calling
  • Streaming output (SSE)
  • Standard OpenAI params: temperature, top_p, max_tokens, etc.
  • Responses endpoint

Not Supported

  • Fine-tuning
  • Files management
  • Organization management
  • Billing management

​
Choose a Model

Not sure which model to use? These two pages are kept up to date with pricing, capability comparisons, and recommendations:

Text / Multimodal Models

Capabilities, pricing, and selection guidance for GPT, Claude, Gemini, Grok, DeepSeek, Qwen, Kimi, GLM, and more.

Image / Video Models

Image models like Nano Banana, GPT-image, Seedream, and Flux, plus video models like VEO, Sora, and Wan — pricing and usage.

​
Basic Information

​
API Endpoints

  • Primary: https://api.apiyi.com/v1
  • Backup: https://vip.apiyi.com/v1

​
Authentication

Every request must include your API Key in the header: 
Authorization: Bearer YOUR_API_KEY

​
Request Format

  • Content-Type: application/json
  • Encoding: UTF-8
  • Method: POST for most endpoints

​
Quick Start

​
Get an API Key

  1. Visit the APIYI console and log in
  2. On the token management page, click “Add” to create an API Key
  3. Copy the generated key for use in your requests

​
Get Multi-Language Code Examples

The console has built-in, ready-to-run code examples for many languages, updated in sync with the latest API version — use these first:
  1. Go to the token management page
  2. On the row of the target API Key, click the 🔧 wrench icon in the “Actions” column
  3. Select “Request Example” to view complete examples in cURL, Python, Node.js, Java, C#, Go, PHP, Ruby, and more
APIYI token management - request examples

​
Online Testing (Playground)

The “API Reference” section provides an online Playground: enter your API Key to send requests and view live responses directly — no code required.

Chat Completions

POST /v1/chat/completions — the main chat and multimodal endpoint.

List Models

GET /v1/models — query currently available models.

Embeddings

POST /v1/embeddings — text vectorization.
Playgrounds for image and video generation endpoints live on their respective model pages (see the image / video model page under “Choose a Model” above).

​
Minimal Example

The most common endpoint — Chat Completions — copy and run. For more parameters and languages, use the Playground above or the console’s “Request Example”: 
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.apiyi.com/v1"
)

response = client.chat.completions.create(
    model="gpt-5-chat-latest",
    messages=[
        {"role": "system", "content": "You are a helpful AI assistant."},
        {"role": "user", "content": "Hello! Please introduce yourself."}
    ],
    temperature=0.7,
    max_tokens=1000
)

print(response.choices[0].message.content)

​
Streaming Response

Set stream: true in the request, and the response is returned chunk by chunk as Server-Sent Events (SSE) — ideal for typewriter-style output: 
stream = client.chat.completions.create(
    model="gpt-5-chat-latest",
    messages=[{"role": "user", "content": "Tell a short joke"}],
    stream=True
)

for chunk in stream:
    content = chunk.choices[0].delta.content or ""
    print(content, end="", flush=True)
Each SSE line starts with data: , and the final line data: [DONE] signals the end.

​
Error Handling

Endpoints follow the OpenAI error format: 
{
  "error": {
    "message": "Invalid API key provided",
    "type": "invalid_request_error",
    "code": "invalid_api_key"
  }
}
Common error codes:
Error CodeHTTP StatusDescription
invalid_api_key401Invalid API key
insufficient_quota429Insufficient balance
model_not_found404Model does not exist
invalid_request_error400Invalid request parameters
rate_limit_exceeded429Request rate too high
server_error500Internal server error
Implement exponential backoff: on 429 / 500, retry with doubling intervals to greatly improve stability. Store your API Key in environment variables — never hard-code it.

​
Rate Limits

Limit TypeDefaultDescription
RPM (requests per minute)3000Per API key
TPM (tokens per minute)1000000Per API key
Concurrent requests100Requests processed simultaneously
Exceeding limits returns 429. Please control your request rate accordingly.

​
Need Help?

Choose a Model

Text / multimodal model recommendations and pricing.

Test Online

Open the API Reference Playground and send requests directly.