Skip to main content
This guide takes you from a new account to a verified model response.
API base URL: https://api.hilinkup.com/v1

1. Create an API key

1

Sign in

Open HiLinkup and sign in.
2

Open API key management

Go to the API key page.
3

Create a dedicated key

Name it after the application and environment, such as local-development. Copy it immediately and store it in a password manager or secret manager.
Never commit an API key, send it in chat, or expose it to browser code.

2. Save the key

3. List available models

Example response:
The response identifies each provider in owned_by and returns a provider-qualified id, such as aliyun/kimi-k3. Copy the full data[].id value to select that provider explicitly. If you send only kimi-k3, HiLinkup uses the platform’s current default provider setting. The examples use aliyun/kimi-k3; replace it if that ID is not returned for your key.
HTTP 200 and a non-empty data array confirm that authentication and model access work.

4. Make a Chat Completions request

Most models currently available on HiLinkup support the OpenAI-compatible Chat Completions protocol first. Use this protocol for the initial connection test.

5. Verify the response

A successful response has:
  • HTTP status 200
  • At least one item in choices
  • Text in choices[0].message.content
  • A finish_reason, normally stop
  • Token counts in usage

Troubleshooting

Check the Bearer token, environment variable, copied key, and whether the key was revoked.
Check account status and model permissions. List models again and select an ID returned for this key.
The complete endpoint is https://api.hilinkup.com/v1/chat/completions. The SDK base URL is https://api.hilinkup.com/v1.
Reduce concurrency, check balance and limits, then retry with bounded exponential backoff.
Copy the exact data[].id from /v1/models; do not use a display name from a screenshot or another account.
This quickstart uses Chat Completions because most currently available models support it first. For a Responses-capable agent or client, see Responses API. The request body, output fields, and streaming events differ.

Migrate from OpenAI

Reuse an existing OpenAI SDK integration.

Connect an application

Configure developer tools, clients, and frameworks.
Last modified on August 12, 2026