Reference documentation. Available models, billing and routing depend on the current site configuration.

FIRST SUCCESSFUL CALL

Verify one request with one key

The integration is complete when the request returns HTTP 200 and the same request appears in 懂api usage logs.

Last updated: 2026-10-17

On this page

STEP 1

Create an API key

Open API keys and choose Create key.
Use Models to confirm which group supports the target model.
Copy the complete sk-... value and keep it out of screenshots and support messages.

STEP 2

Choose the Base URL

Task	Base URL	Use it for
Standard	`http://localhost:3000/v1`	Text, streaming, Responses, embeddings, and tools
Direct	`http://localhost:3000/v1`	Images, videos, long-running requests, and gateway timeouts
Anthropic clients	`http://localhost:3000`	Clients that append `/v1/messages`

Base URL is not the complete request URL

Client settings usually stop at /v1. Append /chat/completions or another resource path only when sending HTTP directly.

STEP 3

Get an available model ID

Copy the exact ID from Models, or list the models available to the current key.

List models

export DONGAPI_API_KEY="sk-your-key"

curl http://localhost:3000/v1/models \
  -H "Authorization: Bearer $DONGAPI_API_KEY"

STEP 4

Send a non-streaming request

Replace YOUR_MODEL_ID with an exact ID returned by the catalog.

POST /v1/chat/completions

curl http://localhost:3000/v1/chat/completions \
  -H "Authorization: Bearer $DONGAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "YOUR_MODEL_ID",
    "messages": [{"role": "user", "content": "Reply with ok"}],
    "stream": false
  }'

Change one variable at a time

Keep the original status and error body, then check the key, URL, model, and group in that order.

ACCEPTANCE

Verify the result

The HTTP status is 200 and the response contains model output.
Usage logs contains a request at the same time with the same model ID.
The log does not show exhausted retries or an upstream failure.