Render a Template and Send an Email Nodes¶

The Render a Template and Send an Email nodes both use Jinja2 templates to produce dynamic text from pipeline state. This page is the complete reference for the shared template variable context, field behavior, and syntax examples for both nodes.

For a plain-English introduction to these nodes, see Node Types.

Shared Template Variable Context¶

Both nodes evaluate templates against the same context dictionary. All keys listed below are available in every template field.

Key	Type	Description
`input`	String	The current input to the node
`node_inputs`	List of strings	All inputs to the node — same as `[input]` except in parallel branches
`temp_state`	Dict	Temporary state for this pipeline run — reset on each new message
`session_state`	Dict	State scoped to the participant's current session
`participant_details`	Dict	Contains `identifier` and `platform` for the current participant
`participant_data`	Dict	Persistent key-value data stored against the participant
`participant_schedules`	List	Scheduled messages associated with the participant

Session-dependent keys

participant_details, participant_data, and participant_schedules are only populated when the pipeline is running inside an active session. They are empty when the pipeline is tested in isolation without a session context.

Render a Template Node¶

Template Field¶

The template field is a single Jinja2 template string. The node renders it against the context and passes the result to the next node.

Example — building a prompt for a downstream LLM node:

You are helping {{ participant_details.identifier }}.

Their question: {{ input }}

Relevant context from their profile:
- Preferred language: {{ participant_data.language | default('English') }}
- Account tier: {{ participant_data.tier | default('standard') }}

Syntax Validation¶

The template string is validated for correct Jinja2 syntax when you save the node. An invalid template prevents saving and displays the line and column of the error.

Send an Email Node¶

Fields¶

Field	Required	Accepts Jinja2	Notes
`recipient`	Yes	Yes	Comma-separated email addresses or a Jinja2 expression that resolves to them
`subject`	Yes	Yes
`body`	No	Yes	When left blank, the node input is used as the body

The node output is always the unmodified node input, regardless of the email content. This means you can insert the node anywhere in a pipeline without breaking the data flow.

Recipient Field Syntax¶

The recipient field must resolve to a comma-separated list of valid email addresses after rendering.

Use case	Template
Single static address	`ops@example.com`
Single address from participant data	`{{ participant_data.email }}`
Multiple addresses from a list	`{{ participant_data.emails \\| join(',') }}`
Addresses stored as a delimited string	`{{ participant_data.emails \\| split(';') \\| join(',') }}`
Mix of static and dynamic	`ops@example.com,{{ participant_data.manager_email }}`

Body Field Behavior¶

When the body field is blank, the node sends the pipeline input as the email body. This is the backwards-compatible default and is appropriate when an upstream node (such as an LLM node) has already produced the text you want to send.

When the body field contains a template, the node renders it and uses the result as the email body. The pipeline input is still passed through to the next node unchanged.

Example — Dynamic Subject and Personalized Body¶

Subject:

Update for {{ participant_details.identifier }}

Body:

Hello {{ participant_data.name | default('there') }},

Here is your update: {{ input }}

To unsubscribe, reply STOP.

Node Types — plain-English overview of all pipeline nodes
Parallel Branches — how node_inputs behaves when multiple branches converge
Participant Data — what is stored in participant_data and how to manage it