v2.5.11

A new PerplexitySearch toolkit gives agents access to the Perplexity Search API, returning ranked web results with titles, URLs, snippets, and publication dates in a single tool call. Built-in filtering by recency and domain makes it straightforward to build agents that need up-to-date, source-controlled retrieval without additional post-processing.

Check out this example of basic search:

from agno.agent import Agent
from agno.tools.perplexity import PerplexitySearch

agent = Agent(tools=[PerplexitySearch()], markdown=True)
agent.print_response("What are the latest developments in AI?")

‍


Details:

• search and asearch (async) functions return a JSON array of results with URL, title, snippet, and date per result
• search_recency_filter restricts results to content from the past day, week, month, or year
• search_domain_filter limits results to a specific list of domains (e.g., reuters.com, bloomberg.com)
• search_language_filter accepts ISO language codes for language-scoped retrieval
• max_results (default 5) and max_tokens_per_page (default 2048) give fine-grained control over result volume and content length
• Requires a PERPLEXITY_API_KEY environment variable; no other configuration needed

‍

See the Perplexity docs for reference.