Exa Tools

Search, answer, and find similar pages with ExaTools, restricting results to specific news domains.

exa_tools.py
"""
Exa Tools
=============================

Demonstrates exa tools.
"""

from agno.agent import Agent
from agno.tools.exa import ExaTools

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------


# Example 1: Enable all tools
agent_all = Agent(
    tools=[
        ExaTools(
            all=True,  # Enable all exa tools
            show_results=True,
        )
    ],
    markdown=True,
)

# Example 2: Enable specific tools only
agent_specific = Agent(
    tools=[
        ExaTools(
            enable_search=True,
            enable_answer=True,
            enable_get_contents=False,
            enable_find_similar=False,
            enable_research=False,
            include_domains=["cnbc.com", "reuters.com", "bloomberg.com"],
            show_results=True,
            text=False,
        )
    ],
    markdown=True,
)

# Example 3: Default behavior (most functions enabled by default)
agent = Agent(
    tools=[
        ExaTools(
            include_domains=["cnbc.com", "reuters.com", "bloomberg.com"],
            show_results=True,
            text=False,
        )
    ],
    markdown=True,
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    agent.print_response("Search for AAPL news", markdown=True)

    agent = Agent(
        tools=[
            ExaTools(
                show_results=True,
            )
        ],
        markdown=True,
    )

    agent.print_response("Search for AAPL news", markdown=True)

    agent.print_response(
        "What is the paper at https://arxiv.org/pdf/2307.06435 about?", markdown=True
    )

    agent.print_response(
        "Find me similar papers to https://arxiv.org/pdf/2307.06435 and provide a summary of what they contain",
        markdown=True,
    )

    agent.print_response(
        "What is the latest valuation of SpaceX?",
        markdown=True,
    )

The first run uses the domain-restricted, text=False configuration. The script then replaces agent with a new, unrestricted toolkit for the remaining prompts, including the arXiv paper. Domain filtering and omission of page text therefore do not apply to those later calls. show_results=True logs tool results; it does not require the final model answer to quote them verbatim.

Run the Example

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate

Install dependencies

uv pip install -U agno exa-py openai

Export your API keys

export EXA_API_KEY="your_exa_api_key_here"
export OPENAI_API_KEY="your_openai_api_key_here"

Run the example

Save the code above as exa_tools.py, then run:

python exa_tools.py

Full source: cookbook/91_tools/exa_tools.py