How to connect SillyTavern to an unlimited, uncensored API

Before you start

• An API keyfrom API keys in your Unquant workspace.
• A recent SillyTavernwith the Custom (OpenAI-compatible) chat completion source.
• Your model IDglm-5.3-flash for GLM-5.3-Flash or qwen-3.8-flash-next for Qwen 3.8 Flash Next.

1. 1

   Open API Connections

   Click the plug icon at the top of SillyTavern to open API Connections.

2. 2

   Choose the connection type

   API Connections

   APIChat Completion

   Chat Completion SourceCustom (OpenAI-compatible)
3. 3

   Enter the endpoint, key and model

   Custom (OpenAI-compatible)

   Custom Endpoint (Base URL)https://api.unquant.io/v1

   Custom API KeyYour Unquant API key

   Model IDglm-5.3-flash

   or, on a Qwen planqwen-3.8-flash-next

   Press Connect. The status turns green when the key is accepted.

4. 4

   Match the context size to your plan

   Open AI Response Configuration (the sliders icon). SillyTavern trims your history to the Context Size set there, so a small value throws away memory you are paying for.

   Context Size

   GLM-5.3-Flash, every planup to 262,144 tokens

   Qwen Roleplay65,536 tokens

   Qwen Plus131,072 tokens

   Qwen Standard and above262,144 tokens
5. 5

   Streaming and reasoning

   • Streamingturn it on so replies appear as they are written.
   • Reasoningboth models think before they answer and return that thinking separately. Enable Request model reasoning to see or hide it, and set the reasoning effort lower, or off, for snappier turns (reasoning controls).

Why long roleplay works better here

Three things break long campaigns: refusals mid-scene, a small window that forgets act one, and a bill that grows with every message because the whole history is re-sent each turn. Here the models are uncensored in the weights, the context goes up to 262K, and every plan is unlimited: re-sending a 200K-token history a thousand times costs the same flat monthly price as sending it once.

Troubleshooting