NextChat custom API setup guide

NextChat, formerly ChatGPT-Next-Web, is a popular open-source AI chat frontend. Configure its hosted or desktop version to use the qionggeme API.

Quick configuration

OPENAI_API_KEY
sk-your API key
BASE_URL
https://qionggeme.com
Use the API host, not the content-site /docs URL.
Recommended access control
CODE=your-password
Example custom models
+claude-sonnet-4-6,+gpt-5.4

NextChat, formerly ChatGPT-Next-Web, has two common deployment modes with different configuration paths:

Create your qionggeme API key on the console token page before continuing.

Setup steps

  1. Self-hosted option 1: set Vercel environment variables

    After forking NextChat and creating the Vercel deployment, add these two required variables:

    OPENAI_API_KEY=sk-your API key
    BASE_URL=https://qionggeme.com

    BASE_URL replaces the default official endpoint. Use the API host, not the content site https://qionggeme.com/docs.

  2. Self-hosted option 2: run NextChat with Docker

    docker run -d -p 3000:3000   -e OPENAI_API_KEY=sk-your API key   -e BASE_URL=https://qionggeme.com   -e CODE=your-access-password   yidadaa/chatgpt-next-web

    CODE protects the app from unauthorized use; separate multiple passwords with commas. Open http://localhost:3000 after the container starts.

  3. Desktop option: configure the client settings

    If you use a NextChat desktop client:

    1. Open Settings with the gear icon in the lower-left corner.
    2. Find the API or OpenAI API settings.
    3. Enter sk-your API key in Open API Key.
    4. Enter https://qionggeme.com in API URL or OpenAI Proxy.
    5. Save, choose a model, and start a conversation.
  4. Optionally customize the visible model list

    NextChat may show only a subset of OpenAI models by default. To add Claude, Grok, and other available IDs, set CUSTOM_MODELS:

    CUSTOM_MODELS=+claude-sonnet-4-6,+claude-opus-4-8,+gpt-5.4,+grok-4.5

    A + adds a model and a - hides one. Use exact IDs from the model pricing page.

Configuration example

# Core NextChat environment variables for Docker or Vercel
OPENAI_API_KEY=sk-your API key
BASE_URL=https://qionggeme.com
CODE=your-password          # recommended access password
CUSTOM_MODELS=+claude-sonnet-4-6,+gpt-5.4

Troubleshooting

NextChat reports Unauthorized or Invalid API Key

Check that the key is complete, BASE_URL equals https://qionggeme.com without /docs, and the account has sufficient balance. A trailing slash is usually unnecessary.

NextChat still sends requests to the official endpoint

The new environment variable was not applied. Redeploy the Vercel project, recreate the Docker container, or configure the API URL separately in the desktop client settings.

Claude or Grok is missing from the model list

Add the exact ID with CUSTOM_MODELS or use the client’s custom-model setting. Compare it with the pricing page.

A public NextChat deployment is consuming my API balance

Set CODE to require an access password, rotate any exposed API key, and avoid embedding secrets in public source code or client-side configuration.

Ready to get started?

Create a qionggeme account, generate an API key, and access every currently available model.

Create an account

Frequently asked questions

Why does NextChat show Unauthorized or Invalid API Key?

Check that the API key is complete and BASE_URL is https://qionggeme.com, not a /docs URL. A trailing slash is usually unnecessary.

Why does NextChat still call the official service after I set BASE_URL?

The environment variable has not been applied. Redeploy on Vercel, recreate the Docker container, or configure the API address separately in the desktop client.

Why is Claude missing from the model list?

Use CUSTOM_MODELS to add a claude- model ID, or add the exact custom model ID in the client’s model settings.

Where does NextChat store conversations?

Conversation data is usually stored locally in the browser using IndexedDB and is not uploaded to the qionggeme server as application history.