Pubmed

Search PubMed for articles with PubmedTools' search_pubmed function, returning expanded article detail when results_expanded is set.

PubmedTools enable an Agent to search PubMed for articles.

Prerequisites

The following example requires the openai library.

uv pip install -U agno openai

Example

The following agent will search PubMed for articles related to "ulcerative colitis".

cookbook/91_tools/pubmed_tools.py
from agno.agent import Agent
from agno.tools.pubmed import PubmedTools

agent = Agent(tools=[PubmedTools()])
agent.print_response("Tell me about ulcerative colitis.")

Toolkit Params

ParameterTypeDefaultDescription
emailstr"your_email@example.com"Specifies the email address to use.
max_resultsOptional[int]NoneOptional parameter to specify the maximum number of results to return.
results_expandedboolFalseReturn expanded article details including author, journal, DOI, URLs, keywords, and MeSH terms.
enable_search_pubmedboolTrueEnable the search_pubmed functionality.
allboolFalseEnable all functionality.

Toolkit Functions

FunctionDescription
search_pubmedSearches PubMed for articles based on a specified query. Parameters include query for the search term and max_results for the maximum number of results to return (default is 10). Returns formatted article summaries. With short abstracts, the default output can omit title and date; use results_expanded=True for expanded citation fields.

Developer Resources