Cache Tool Calls

Combine cache_results and stop_after_tool_call on a @tool-decorated function.

cache_tool_calls.py
"""
Cache Tool Calls
=============================

Demonstrates cache tool calls.
"""

import json

import httpx
from agno.agent import Agent
from agno.tools import tool

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


@tool(stop_after_tool_call=True, cache_results=True)
def get_top_hackernews_stories(num_stories: int = 5) -> str:
    # Fetch top story IDs
    response = httpx.get("https://hacker-news.firebaseio.com/v0/topstories.json")
    story_ids = response.json()

    # Yield story details
    stories = []
    for story_id in story_ids[:num_stories]:
        story_response = httpx.get(
            f"https://hacker-news.firebaseio.com/v0/item/{story_id}.json"
        )
        story = story_response.json()
        if "text" in story:
            story.pop("text", None)
        stories.append(json.dumps(story))

    return "\n".join(stories)


agent = Agent(
    tools=[get_top_hackernews_stories],
    markdown=True,
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    agent.print_response("What are the top hackernews stories?", stream=True)

Repeat the same tool call with the same arguments to exercise a cache hit; this script makes one agent request. The default cache lifetime is 3,600 seconds, with files under the temporary directory’s agno_cache folder. Cached headlines can therefore be up to an hour old. stop_after_tool_call=True returns the tool output without another model response.

Run the Example

Set up your virtual environment

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

Install dependencies

uv pip install -U agno openai

Export your OpenAI API key

export OPENAI_API_KEY="your_openai_api_key_here"

Run the example

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

python cache_tool_calls.py

Full source: cookbook/91_tools/tool_decorator/cache_tool_calls.py