NextChat with Nbility
NextChat’s self-hosted endpoint configuration differs from a typical SDK: the official README defines BASE_URL as the OpenAI service origin. Do not automatically add /v1.
Self-hosted configuration
Set these in the deployment environment or a private environment file:
OPENAI_API_KEY=YOUR_API_KEY
BASE_URL=https://api.nbility.ai
CUSTOM_MODELS=+YOUR_MODEL_ID
Replace the key and model ID, then restart or redeploy. CUSTOM_MODELS adds the model to the picker; merge existing entries instead of replacing them.
Keep the key in server-side configuration. For a shared instance, configure access controls according to NextChat’s deployment guide so visitors cannot freely spend a personal API balance.
Client settings
Versions allowing custom user settings can accept your key and model in the UI. If the address follows the same origin rule as the server, use https://api.nbility.ai. If that version explicitly asks for an OpenAI base ending in /v1, follow its field documentation.
The final request should be /v1/chat/completions. For 404 errors, check duplicated /v1 rather than inferring path behavior from the field name alone.
Verify
Start a new conversation, select the added model, send a short message, and inspect Usage logs. Existing conversations or masks may retain separate model selections.
References
Adapted and rewritten for Nbility from aicentos’s NextChat guide, licensed under CC BY 4.0. Its account, plan, and proprietary service instructions do not apply here. See the client’s official documentation.