🔌 YingCore Integration Guide

👤 Best for: Developers and teams who need to plug the platform into their own systems
⏱️ Reading time: about 8 minutes
💡 In one sentence: Use the OpenAI protocol to call every configured model through one unified API.

YingCore exposes an OpenAI-compatible API. You can integrate with any language or SDK that supports the OpenAI protocol, and call every model that is configured on the platform.

1. Get an API Key

1. Log in to the platform → click "Get API" at the top.
2. In the API Key area, click "Create" to generate a platform token.
3. Copy and save the Key and the Base URL:
   • Base URL: https://your-platform-url/platform-api/v1
   • API Key: YOUR_API_KEY

The Key is only fully visible when it is created — save it right away. Treat the Key as a secret credential.

2. List Available Models

Before calling, fetch the model list to confirm which model IDs are available:

GET https://your-platform-url/platform-api/v1/models
Authorization: Bearer YOUR_API_KEY

The response includes each model ID, vendor, type, context length, and capability tags. Use a model ID from the list as the model field in chat requests.

3. Code Samples in Four Languages

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://your-platform-url/platform-api/v1",
    api_key="YOUR_API_KEY",
)
resp = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

cURL

curl https://your-platform-url/platform-api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-chat","messages":[{"role":"user","content":"Hello"}]}'

Node.js

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://your-platform-url/platform-api/v1",
  apiKey: "YOUR_API_KEY",
});

const resp = await client.chat.completions.create({
  model: "deepseek-chat",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(resp.choices[0].message.content);

Java

import com.theokanning.openai.OpenAiService;
import com.theokanning.openai.completion.chat.*;

// baseUrl points to the platform
OpenAiService service = new OpenAiService("YOUR_API_KEY", Duration.ofSeconds(60));
// Configure baseUrl: https://your-platform-url/platform-api/v1

All four languages integrate through the official OpenAI SDK — you only need to change base_url and api_key. The exact class names depend on the SDK version you use.

4. Push API Config to a Client

The Get API page also offers a Push API action so any registered client (such as YingClaw) can use platform models in one click:

1. In the Push API area, pick the target client device (filter by user email).
2. Pull the global / user configuration.
3. Click "Quick Fill" to auto-populate:
   • api_key — the Key assigned to the client.
   • default_provider — custom:http://122.114.233.81/platform-api/v1/chat/completions (subject to your deployment).
   • default_model — for example deepseek-chat.
4. Click Push to Client — the client is now wired up.

5. Embed on an External Site

To embed the platform chat assistant into your own website:

1. In the Workbench, open the app card menu and choose "Embed on Site".
2. Generate the iframe embed code (/chat/share?shared_id=...) — you can hide the avatar, set the language, and so on.
3. Paste the code into your web page.

Some embed modes (partial embed / browser extension) are marked as "coming soon".

6. Integration Notes

• Protocol: Compatible with the OpenAI protocol — minimal changes to existing OpenAI code.
• Authentication: Request header Authorization: Bearer <API_KEY>.
• Usage: View call counts and Token consumption under Get API → Usage Statistics; admins can view usage per user or tenant in YingClaw.
• Rate limits: If you hit a rate limit or quota issue, contact the platform admin to adjust.
• Model availability: A model is callable only if it has been configured and enabled in Model Providers. Unconfigured models cannot be called.