Perplexity AI search-grounded chat, search, and embeddings
Perplexity is an AI answer engine that runs live web searches and returns cited, up-to-date answers. The NoClick Perplexity node calls the Sonar API, so a single workflow can run chat completions grounded in current web results, search the web directly, pull academic sources, and query SEC filings, plus force clean structured JSON out of any prompt. It also creates embeddings for search and similarity, and can fire long-running requests as async chat completions you retrieve later. Add it to a canvas when you want research and grounded answers to run on a schedule or in response to another node instead of a person typing into a search box.
Yes. The Search the Web operation and the grounded chat completions query current web data and return answers with citations, so you get fresh information rather than a model's stale training data. You can drop the answer and its sources into any downstream node.
Authenticate with a Perplexity Sonar API key, which you paste into the node's credential field. The node also exposes Generate Auth Token and Revoke Auth Token operations for managing tokens programmatically once you are connected.
The Perplexity node is request and response, so it does not fire on its own. Wire it to any trigger in NoClick, such as a schedule, a form submission, or an inbound message, and that event will kick off the search or chat completion automatically.
Yes. The Structured JSON Output operation makes the model answer as a defined JSON object rather than prose, so fields like category, sentiment, or extracted values come back ready to route or store. That makes it clean to chain into later nodes without parsing free text.
They handle long-running or heavy requests without blocking a workflow. Fire Create Async Chat Completion to start the job, then use Get Async Chat Completion to fetch the result later or List Async Chat Completions to see what is in flight.