Skip to content
Gonka Broker

Connect SillyTavern to Gonka Broker

SillyTavern is a self-hosted frontend for chatting with LLMs, and the most popular tool for character and roleplay chats. It speaks OpenAI-compatible APIs natively, so you can point it at Gonka Broker and run your chats on open-source models.

Prerequisites

Section titled “Prerequisites”

Connect Gonka Broker

Section titled “Connect Gonka Broker”
  1. Open the API Connections panel (the plug icon in the top bar).
  2. Set API to Chat Completion.
  3. Set Chat Completion Source to Custom (OpenAI-compatible).
  4. Fill in:
    • Custom Endpoint (Base URL): https://proxy.gonkabroker.com/v1
    • Custom API Key: your Gonka key (gnk-prx-…)
  5. Click Connect. The Available Models dropdown fills with Gonka Broker‘s live model list.
  6. Pick a model, for example MiniMaxAI/MiniMax-M2.7 (see Supported Models).

See the model’s thinking (optional)

Section titled “See the model’s thinking (optional)”

The models Gonka Broker serves are reasoning models: they think before they answer. The thinking arrives in a separate reasoning response field, so your chat text is always clean — nothing to configure. If you want to see the reasoning:

  1. Open the AI Response Configuration panel (the sliders icon in the top bar).
  2. Enable Request model reasoning.

SillyTavern reads the reasoning field natively for the Custom source and folds it into a collapsible Thinking block above each reply — tidy chats, and you can peek at the reasoning when you want to.

Verify

Section titled “Verify”

Send a message in any chat. A reply means SillyTavern is reaching Gonka Broker through your key: clean answer text, with the thinking available as a collapsible block if you enabled Request model reasoning.

Troubleshooting

Section titled “Troubleshooting”
  • Available Models stays empty after Connect: confirm Custom Endpoint (Base URL) is exactly https://proxy.gonkabroker.com/v1 (including the /v1) and the key is active.
  • 401 / invalid API key: wrong or paused key. Create a fresh one from Create a Gonka API Key.
  • No Thinking block above replies: enable Request model reasoning in AI Response Configuration (see above).