Chatbot UI with Nbility

Custom endpoint support differs across Chatbot UI’s current branch, older versions, and community forks. Setting OPENAI_API_HOST is not a universal solution: the current main-branch environment example does not list that field.

Check your instanceCopy link

Identify the repository and release, then look for custom OpenAI-compatible models in settings. You need editable URL, key, and model ID fields; an OpenAI key field alone is insufficient.

For a release with custom model management:

1. Create a Nbility model connection.
2. Select OpenAI Chat Completions.
3. Use https://api.nbility.ai/v1 for a base URL, or https://api.nbility.ai/v1/chat/completions when a full endpoint is explicitly required.
4. Supply a dedicated key and model ID, save, and select it in a new conversation.

Self-hosted deploymentCopy link

Follow the project’s own authentication and data-storage setup. A custom model connection does not replace dependencies such as Supabase.

Only set an endpoint environment variable when your deployed version explicitly reads it. Restart after changing it. Do not transfer variable names between forks or put keys in NEXT_PUBLIC_ variables.

Verify or choose an alternativeCopy link

Send a short message and inspect Usage logs. If the branch lacks custom endpoints, use NextChat or LobeChat.

ReferencesCopy link

Adapted and rewritten for Nbility from aicentos’s Chatbot UI guide, licensed under CC BY 4.0. Its account, plan, and proprietary service instructions do not apply here. See the client’s official documentation.