Connect n8n to Gonka Broker

n8n is a workflow-automation tool with built-in AI nodes. Those nodes speak the OpenAI API, so you can point them at Gonka Broker with a custom Base URL on the OpenAI credential and run your automations on open-source models.

You’ll connect through the AI Agent node and its OpenAI Chat Model sub-node. (The standalone OpenAI app node is best avoided here — its v2 has a known runtime issue with custom base URLs.)

Prerequisites

• A Gonka API key (starts with gnk-prx-). See Create a Gonka API Key.
• A running n8n instance (n8n Cloud or self-hosted).

Create the OpenAI credential

1. In the left sidebar, click the Create button at the top and choose Credential (or open Credentials from the sidebar and click Add credential).
2. In the Add new credential search box, type OpenAI, select OpenAI (the OpenAi API type), and click Continue.
3. Fill in:
   • API Key — your Gonka key (gnk-prx-…)
   • Base URL — https://proxy.gonkabroker.com/v1
   • Organization ID — leave blank
4. Save. n8n runs a credential test against the Base URL; a green check means it reached Gonka Broker.

Add the model to a workflow

1. In a workflow, add an AI Agent node (or any AI chain node).
2. On its Chat Model input, add the OpenAI Chat Model sub-node and select the credential you just created.
3. Open the Model dropdown — n8n loads the list from Gonka Broker‘s /v1/models. Pick moonshotai/Kimi-K2.6 (recommended) or MiniMaxAI/MiniMax-M2.7 (see Supported Models).
4. Turn off the Use Responses API toggle on the OpenAI Chat Model node. It is on by default and routes requests to OpenAI’s Responses API endpoint, which Gonka Broker does not serve — leaving it on fails with a “resource not found” (404) error at runtime. With it off, the node uses the standard chat-completions endpoint.

Verify

Run the workflow (or use the Chat trigger / Test workflow) and send a message. A reply from the AI Agent confirms n8n is reaching Gonka Broker through your key.

Troubleshooting

• “The resource you are requesting could not be found” (404) at runtime — the OpenAI Chat Model node’s Use Responses API toggle is on (the default); turn it off so the node calls the chat-completions endpoint Gonka Broker serves. Also prefer the AI Agent + OpenAI Chat Model node over the standalone OpenAI app node (its v2 has a separate custom-base-URL runtime bug).
• 401 / invalid API key — wrong or paused key. Create a fresh one from Create a Gonka API Key.
• Model list empty or model errors — pick an id Gonka Broker actually serves: moonshotai/Kimi-K2.6 or MiniMaxAI/MiniMax-M2.7 (see Supported Models).
• Visible <think> reasoning in the output — reasoning models such as MiniMaxAI/MiniMax-M2.7 emit their thinking inline, which n8n shows verbatim. Use moonshotai/Kimi-K2.6 for clean output.