Agent answer through a template
The model's answer passes through a Jinja template — one consistent shape for every reply.
Graph preview
How it works
A Template node sits after the agent and drops the model’s answer into a fixed text via
{{ nodes.agent.output }}. The presentation stops depending on the model’s mood — heading,
signature and markup are set by the graph, not by the prompt.
- Entry → Agent — the model answers.
- Template wraps that answer in a constant shape.
- Answer returns the finished text.
What to configure after import
- Edit the template text:
{{ nodes.<id>.output }},{{ inputs.<port> }}and{{ variables.<name> }}are available — see templates. - Templates are also the place for disclaimers: a line like “generated by AI” ends up under every answer, and the model cannot forget it.