Zihin Chat Model (Deprecated)
AI-generatedSummary
Provides a Zihin chat model to connect to an n8n AI Agent. Select a model or use Auto to route by task type; the node is deprecated and the notice recommends using Zihin Agent instead.
Inputs
- Model Name or ID (required) — Choose a listed model or provide its ID. Auto uses smart routing; its optional Task Type guides model selection.
- Options — Configure generation settings such as temperature, token limit, sampling penalties, stop sequences, timeout, retries, and session ID. You can also set a JSON output schema, limit retained context messages, set a message-count limit, adjust telemetry, or override the base URL with the relevant security options.
- Request Options — Configure request batching, SSL certificate handling, an HTTP proxy, or a request timeout.
Output shape
A chat model object supplied through the node's Model output for an AI Agent to use; it does not return a standalone list of response items.
The model sends the agent's chat messages to the Zihin LLM gateway and returns generated text, with tool calls when present and token usage/model metadata. A session ID can be used to keep context across API calls. The node uses a direct gateway outside RBAC, quota, HITL, and audit; removal is planned for v1.0.0.