The Agent Node
Operator-controlled AI — how the Agent node reads the snapshot, calls the provider, and reports back, with nothing hardcoded.
Agent: Provider Credentials
Configuring the per-provider AI API keys the Agent node uses, and how it picks the right one at runtime.
Agent: Slug & Assigned Prompt
How setting the agent slug binds the node to the operator's configured agent — and the prompt that comes with it.
Agent: The Custom Prompt
The runtime input you feed the agent — the subject or data it works on, distinct from its assigned prompt.
Agent: System Context
Optionally injecting the operator's system context into the agent's prompt — all of it, or named entries.
Agent: Auto-Logging & Self-Termination
The two things the Agent node does for you after every AI call — logs the result, and stops the run if it breaches its limits.
Agent: Error Handling & Node Options
How the Agent node behaves when a call fails, and how n8n's On Error and Retry settings change what it does.