Node Reference
Overview of all node types in the Linea workflow engine, with links to per-node documentation.
Node Reference
Every workflow is a directed graph of nodes connected by edges. Each node type has a specific purpose — from running LLM inference to sending Slack messages to branching based on a condition.
Variable Substitution
All string fields in node configuration support {{variableName}} substitution:
| Syntax | Resolves to |
|---|---|
{{input.field}} | Field from the trigger input payload |
{{nodeName}} | Output of a previous node (by name or ID) |
{{lastOutput}} | Output of the most recently completed node |
{{nodeResults.nodeId.output}} | Deep-access a specific node result |
{{variables.myVar}} | Any variable set during the current execution |
Substitution is performed recursively — nested references like {{input.data.items.0.name}} work as expected.
Variable substitution is available in all string fields of all node types. Object fields (e.g. httpHeaders entries) have their string values substituted individually.
Node Categories
Core
| Node | Description |
|---|---|
| Start | Entry point — initializes state from trigger input |
| End | Terminal node — marks execution complete |
AI
| Node | Description |
|---|---|
| Agent | LLM node with tool use, structured output, and long-term memory |
| Memory | Read or clear in-workflow key-value memory |
Logic
| Node | Description |
|---|---|
| If-Else | Branch on a boolean Jexl expression |
| Router | Multi-branch routing via prioritized conditions |
Tools
Data
| Node | Description |
|---|---|
| Transform | Reshape variables with a Jexl expression |
| Extract | Scrape web content via Firecrawl or native fetch |
| Retriever | Search a workspace knowledge base |
| Code | Sandboxed JS execution (currently disabled) |
Flow
| Node | Description |
|---|---|
| Loop | Iterate over an array with optional per-item transform |
| Subworkflow | Call another workflow synchronously |
Safety
| Node | Description |
|---|---|
| Guardrails | PII, jailbreak, and harmful content detection |
| Approval Gate | Human-in-the-loop checkpoint |
Integrations
All integration nodes require the corresponding secret stored in Settings → Credentials.
| Node | Credential | Description |
|---|---|---|
| Slack | SLACK_TOKEN | Send messages or DMs, list channels |
| GitHub | GITHUB_TOKEN | Issues, comments, pull requests |
| Notion | NOTION_TOKEN | Pages, blocks, database queries |
| Gmail | GMAIL_TOKEN | Send and read emails |
Misc
| Node | Description |
|---|---|
| Note | Canvas annotation — no runtime effect |