NagentArtefacts
Dev & Technology
Enterprise and White Glove
Company
Openperplex connects to a Nagent workspace with an API key. Once it is connected, agents can call 8 Openperplex actions, such as "Get Website Markdown", "Get Website Screenshot" and "Get Website Text". Nothing is enabled on connect: each action is allowed one at a time, and an action with side effects runs or waits for a person according to the agent's level.
Every operation an agent can call against Openperplex, with input parameters and output schema.
OPENPERPLEX_CUSTOM_SEARCHTool to perform a custom search with optional filtering and field selection. Use when you need tailored results based on query, filters, size, or fields.
Input parameters
Number of desired results (must be ≥ 1).
The search query string.
Filters to apply on the search results.
User prompt for guiding the search.
System prompt for guiding the search.
Fields to exclude from results.
Fields to include in results.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_CUSTOM_SEARCH_STREAMTool to perform custom streaming search with custom prompts. Use when you need AI-powered search responses with custom system and user prompts for more specific or tailored queries. Returns aggregated text response along with optional sources and images.
Input parameters
Model to use for the search. Options: 'gpt-4o-mini' (default), 'gpt-4o', 'o3-mini-high', 'o3-mini-medium', 'gemini-2.0-flash'.
Top-p for output diversity (0.0-1.0, default 0.9).
Two-letter country code for localized results, e.g., 'us', 'uk', 'fr', 'de', 'jp'.
Type of search: 'general' or 'news'.
Temperature for output randomness (0.0-1.0 exclusive, default 0.2).
User prompt specifying the search query. Be specific as this is used for web search.
Whether to include images in the response.
System prompt defining assistant behavior and response format.
Filter results by recency: 'hour', 'day', 'week', 'month', 'year', or 'anytime'.
Whether to include sources in the response.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_GET_WEBSITE_MARKDOWNTool to retrieve the markdown content of a specified website. Use after confirming the URL. Supports optional formats like 'gfm' or 'commonmark'.
Input parameters
The full URL of the website to fetch and convert to markdown
The target markdown standard/format
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_GET_WEBSITE_SCREENSHOTTool to capture a screenshot of a website. Use after confirming the target URL is reachable.
Input parameters
The website URL to capture.
Whether to capture the entire scrollable page.
Width of the browser viewport in pixels.
Height of the browser viewport in pixels.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_GET_WEBSITE_TEXTTool to retrieve the main text content of a specified website URL. Use when you need content extraction from online articles or pages. Use after confirming the URL is publicly accessible.
Input parameters
The URL of the web page to extract text from.
If true, returns cleaner and more concise text.
Maximum number of characters to extract.
Format of the returned text: 'plain', 'html', or 'markdown'.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_QUERY_FROM_URLTool to query documents from a URL. Use when you need to fetch and interrogate web-hosted content with a natural language question.
Input parameters
The URL of the website you want to interact with.
LLM model to use. Options: o3-mini-high, o3-mini-medium, gpt-4o, gpt-4o-mini, gemini-2.0-flash. Default is gpt-4o-mini.
The question you want to ask about the content at the URL.
Format of the answer. Options: 'text', 'markdown', or 'html'. Default is 'text'.
Language code for the response. 'auto' will auto-detect based on the query. Supported: auto, en, fr, es, de, it, pt, nl, ja, ko, zh, ar, ru, tr, hi.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_SEARCHTool to search documents using query parameters. Use when you have a search query and optional filters ready.
Input parameters
Model to use for generating responses. Options: 'o3-mini-high', 'o3-mini-medium', 'gpt-4o', 'gpt-4o-mini', 'gemini-2.0-flash'.
The search query or question you want to ask. This is the primary input for your search.
Country code for localized search results (e.g., 'us', 'jp', 'br', 'fr').
Format of the answer. Options: 'text', 'markdown', or 'html'.
Type of search to perform. Options: 'general' or 'news'.
Optional date for context. Format example: 'Today is Monday 16 of September 2024 and the time is 6:36 PM'. If not provided, the API uses the current date of the server.
Set to True to include images in the result.
Filter results by recency. Options: 'hour', 'day', 'week', 'month', 'year', 'anytime'.
Set to True to include sources in the result.
Set to True to include citations in the response.
Language code for the response. Use 'auto' to auto-detect based on the query.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
OPENPERPLEX_SEARCH_STREAMTool to stream search results from OpenPerplex. Use when real-time AI-powered search responses are needed. Returns aggregated text response along with optional sources, images, and citations.
Input parameters
Model to use for the search. Options: 'gpt-4o-mini' (default), 'gpt-4o', 'o3-mini-high', 'o3-mini-medium', 'gemini-2.0-flash'.
Search query string.
Two-letter country code for localized results, e.g., 'us', 'uk', 'fr', 'de', 'jp'.
Format of the answer: 'text', 'markdown', or 'html'.
Type of search: 'general' or 'news'.
Optional date context string, e.g., 'Today is Monday 16 of September 2024 and the time is 6:36 PM'. Uses server date if not provided.
Whether to include images in the response.
Filter results by recency: 'hour', 'day', 'week', 'month', 'year', or 'anytime'.
Whether to include sources in the response.
Whether to include citations in the response.
Language code for response. 'auto' auto-detects based on query. Options: en, fr, es, de, it, pt, nl, ja, ko, zh, ar, ru, tr, hi.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
Agents can call 8 Openperplex actions on Nagent, including "Get Website Markdown", "Get Website Screenshot" and "Get Website Text". Get Website Markdown: Retrieve the markdown content of a specified website. Each action is listed on this page with its input parameters and its output.
Openperplex connects with an API key, under your workspace's own connection. Nothing is enabled on connect: each action is allowed one at a time and can be scoped to the agents that need it.
Get Website Markdown takes 1 required input: url. It also takes 1 optional input: markdown_format. It returns data, error and successful.