SillyTavern with Nbility

SillyTavern connects through a custom OpenAI-compatible Chat Completion source. Its settings differ from the original TavernAI.

Create an API connectionCopy link

1. Open API Connections in the top toolbar.
2. Select Chat Completion, then Custom (OpenAI-compatible); wording varies by release.
3. Enter https://api.nbility.ai/v1 as the custom endpoint.
4. Supply a dedicated token and an exact model ID.
5. Save, connect, and select this connection for the character conversation.

Do not choose legacy Text Completion or unrelated protocol sources. The custom source appends /chat/completions; the configured address is not a full request URL.

Set context and response limitsCopy link

Match context and response limits to the actual model. Character cards, world info, history, and system prompts all consume context.

Start with a short card and limited history, then verify streaming and additional supported features.

Verify and troubleshootCopy link

A connection indicator or model list only verifies part of the flow. Send an actual message and check the model and status in Usage logs.

• 404: check whether a full endpoint was entered and the path appended twice.
• Empty model list: enter the ID manually and verify token permissions.
• Empty or malformed replies: start with basic prompts and review reasoning or output settings required by the model.

ReferencesCopy link

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