Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.perplexity.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

SuperPlane is an open-source DevOps control plane for long-lived, event-driven workflows. It exposes Perplexity as a native component on the canvas — drop a Perplexity node into any workflow and it runs a Perplexity agent with web search and URL fetching, then emits the response and source citations as a payload for downstream nodes.
SuperPlane models workflows as event-driven canvases: nodes emit payloads, downstream nodes subscribe to them, and the runtime tracks every run with full observability. Learn more at superplane.com.

​
Component: Perplexity

The Perplexity component runs a Perplexity AI agent as a workflow step. Use it for research, synthesis, automated analysis, and content generation grounded in real-time web sources.

​
Action

ActionDescription
Run AgentRun a Perplexity AI agent with web search and URL fetching capabilities.

​
Configuration

ParameterDescription
PresetAgent preset to use: fast-search, pro-search, deep-research, or advanced-deep-research. When set, Model is ignored.
ModelModel identifier to use when no preset is specified (e.g., sonar-pro).
InputThe prompt or question for the agent. Supports SuperPlane expressions.
InstructionsOptional system-level instructions for the agent.
Web SearchEnable the web_search tool (default: true).
Fetch URLEnable the fetch_url tool (default: true).

​
Output Payload

The component emits a payload with these fields downstream nodes can access via expressions:
FieldDescription
textThe generated text response.
citationsSource citations from web results.
modelThe specific model used for this run.
usageToken and cost usage information.

​
API Key Setup

Configure the Perplexity integration in SuperPlane with your API key. Open your canvas’s Integrations settings, add a Perplexity integration, and paste your key.

Get API Key

Generate your Perplexity API key from the API portal.

​
Quick Start

A minimal workflow that asks Perplexity a question whenever a manual run is triggered:
1

Add a Manual Run trigger

Drop a Manual Run node onto the canvas — this is the workflow’s entry point.
2

Add a Perplexity node

Click + Components, choose Perplexity → Run Agent, and drag it onto the canvas. Connect Manual Run → Perplexity.
3

Configure the agent

In the Perplexity node:
  • Preset: pro-search
  • Input: What is the latest news on nuclear fusion?
  • Web Search: enabled
  • Fetch URL: enabled
4

Run

Click Run on the Manual Run node. The Perplexity node emits a payload with text, citations, model, and usage.

​
Reading the Response Downstream

Use SuperPlane expressions to pipe the Perplexity output into the next node — for example, a Slack message or a database write: 
Research result: {{ $['Perplexity'].data.text }}

Sources:
{{ $['Perplexity'].data.citations }}
In condition fields (If / Filter), write expressions without {{ }}: 
$['Perplexity'].data.usage.total_tokens > 1000

​
Use Cases

  • Incident triage — when an alert fires, run a Perplexity research step to gather background on a vendor, library, or CVE before paging on-call.
  • Release notes synthesis — turn raw GitHub commit diffs into customer-facing release notes with citations.
  • Competitive monitoring — schedule a daily canvas that queries Perplexity for industry news and posts a summary to a Slack channel.
  • Document grounding — feed a Perplexity run’s output into a downstream component that updates a Notion page or knowledge base.

SuperPlane Perplexity Component

Official component documentation.

SuperPlane Docs

Full SuperPlane documentation.

Perplexity Agent API

How the Perplexity agent layer works.

Perplexity Models

Available Sonar models.

​
Support

Need help with the integration?