Perplexity

The Perplexity component is an AI component that allows users to connect the AI models served on the Perplexity Platform. It can carry out the following tasks:

Chat

Release Stage

Alpha

Configuration

The component definition and tasks are defined in the definition.yaml and tasks.yaml files respectively.

Setup

In order to communicate with Perplexity, the following connection details need to be provided. You may specify them directly in a pipeline recipe as key-value pairs within the component's setup block, or you can create a Connection from the Integration Settings page and reference the whole setup as setup: ${connection.<my-connection-id>}.

Field	Field ID	Type	Note
API Key	`api-key`	string	Fill in your API key from the vendor's platform.

Supported Tasks

Chat

Generate response base on conversation input.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_CHAT`
Chat Data (required)	`data`	object	Input data.
Input Parameter	`parameter`	object	Input parameter.

Input Objects in Chat

Chat Data

Input data.

Field	Field ID	Type	Note
Chat Messages	`messages`	array	List of chat messages.
Model Name	`model`	string	The model to be used for `TASK_CHAT`. Enum values `sonar` `sonar-pro` `llama-3.1-sonar-small-128k-online` `llama-3.1-sonar-large-128k-online` `llama-3.1-sonar-huge-128k-online`

Chat Messages

List of chat messages.

Field	Field ID	Type	Note
Content	`content`	array	The message content.
Name	`name`	string	An optional name for the participant. Provides the model information to differentiate between participants of the same role.
Role	`role`	string	The message role, i.e. 'system', 'user' or 'assistant'. Enum values `system` `user` `assistant`

Content

The message content.

Field	Field ID	Type	Note
Text Message	`text`	string	Text message.
Text	`type`	string	Text content type.

Input Parameter

Input parameter.

Field	Field ID	Type	Note
Enable Search Classifier	`enable-search-classifier`	boolean	Whether to enable search classifier.
Frequency Penalty	`frequency-penalty`	number	A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. Incompatible with `presence_penalty`.
Last Updated After Filter	`last-updated-after-filter`	string	Filters search results to only include content last updated after this date. Format should be %m/%d/%Y (e.g. 3/1/2025)
Last Updated Before Filter	`last-updated-before-filter`	string	Filters search results to only include content last updated before this date. Format should be %m/%d/%Y (e.g. 3/1/2025)
Max New Tokens	`max-tokens`	integer	The maximum number of completion tokens returned by the API. The total number of tokens requested in max_tokens plus the number of prompt tokens sent in messages must not exceed the context window token limit of model requested. If left unspecified, then the model will generate tokens until either it reaches its stop token or the end of its context window.
Presence Penalty	`presence-penalty`	number	A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with `frequency_penalty`.
Return Related Questions	`return-related-questions`	boolean	Determines whether related questions should be returned.
Search After Date Filter	`search-after-date-filter`	string	Filters search results to only include content published after this date. Format should be %m/%d/%Y (e.g. 3/1/2025)
Search Before Date Filter	`search-before-date-filter`	string	Filters search results to only include content published before this date. Format should be %m/%d/%Y (e.g. 3/1/2025)
Search Domain Filter	`search-domain-filter`	string	Given a list of domains, limit the citations used by the online model to URLs from the specified domains. Currently limited to only 3 domains for whitelisting and blacklisting. For blacklisting add a `-` to the beginning of the domain string.
Search Mode	`search-mode`	string	Controls the search mode used for the request. When set to 'academic', results will prioritize scholarly sources like peer-reviewed papers and academic journals. When set to 'sec', results will prioritize financial and legal sources. Enum values `academic` `web` `sec`
Search Recency Filter	`search-recency-filter`	string	Returns search results within the specified time interval - does not apply to images. Values include `month`, `week`, `day`, `year`.
Stream	`stream`	boolean	If set, partial message deltas will be sent. Tokens will be sent as data-only server-sent events as they become available.
Temperature	`temperature`	number	The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic.
Top K	`top-k`	number	The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. We recommend either altering top_k or top_p, but not both.
Top P	`top-p`	number	The nucleus sampling threshold, valued between 0 and 1 inclusive. For each subsequent token, the model considers the results of the tokens with top_p probability mass. We recommend either altering top_k or top_p, but not both.
Web Search Options	`web-search-options`	object	Configuration for using web search in model responses.

Web Search Options

Configuration for using web search in model responses.

Field	Field ID	Type	Note
Search Context Size	`search-context-size`	string	Determines how much search context is retrieved for the model. Options are: low (minimizes context for cost savings but less comprehensive answers), medium (balanced approach suitable for most queries), and high (maximizes context for comprehensive answers but at higher cost). Enum values `low` `medium` `high`
User Location	`user-location`	object	To refine search results based on geography, you can specify an approximate user location.

User Location

To refine search results based on geography, you can specify an approximate user location.

Field	Field ID	Type	Note
Country	`country`	string	The two letter ISO country code of the user's location.
Latitude	`latitude`	number	The latitude of the user's location.
Longitude	`longitude`	number	The longitude of the user's location.

Output	Field ID	Type	Description
Output Data	`data`	object	Output data.
Output Metadata (optional)	`metadata`	object	Output metadata.

Output Objects in Chat

Output Data

Field	Field ID	Type	Note
Choices	`choices`	array	List of chat completion choices.
Citations	`citations`	array	List of citations.
Search Results	`search-results`	array	A list of search results related to the response.

Choices

Field	Field ID	Type	Note
Created	`created`	integer	The timestamp of when the chat completion was created. Format is in ISO 8601. Example: 2024-07-01T11:47:40.388Z.
Finish Reason	`finish-reason`	string	The reason the model stopped generating tokens.
Index	`index`	integer	The index of the choice in the list of choices.
Message	`message`	object	A chat message generated by the model.

Message

Field	Field ID	Type	Note
Content	`content`	string	The contents of the message.
Role	`role`	string	The role of the author of this message.

Search Results

Field	Field ID	Type	Note
Date	`date`	string	The date of the search result.
Title	`title`	string	The title of the search result.
URL	`url`	string	The URL of the search result.

Output Metadata

Field	Field ID	Type	Note
Usage	`usage`	object	Usage statistics for the request.

Usage

Field	Field ID	Type	Note
Completion Tokens	`completion-tokens`	integer	Number of tokens in the generated response.
Prompt Tokens	`prompt-tokens`	integer	Number of tokens in the prompt.
Total Tokens	`total-tokens`	integer	Total number of tokens used in the request (prompt + completion).

Example Recipes

version: v1beta

variable:
  prompt:
    type: string
    title: Prompt

component:
  perplexity-0:
    type: perplexity
    task: TASK_CHAT
    input:
      data:
        model: sonar
        messages:
          - content:
            - text: Be precise and concise.
              type: text
            role: system
          - content:
            - text: ${variable.prompt}
              type: text
            role: user
            name: Miles
      parameter:
        max-tokens: 500
        temperature: 0.2
        top-p: 0.9
        stream: false
        search-domain-filter:
          - perplexity.ai
        search-recency-filter: month
        top-k: 0
        presence-penalty: 0
        frequency-penalty: 1

output:
  perplexity:
    title: Perplexity
    value: ${perplexity-0.output}