Add Dependencies to Agent Context

This example demonstrates how to create a context-aware agent that can access real-time HackerNews data through dependency injection, enabling the agent to provide current information.

Create a Python file

1import json
2
3import httpx
4
5from kern.agent import Agent
6from kern.models.openai import OpenAIResponses
7
8
9def get_top_hackernews_stories(num_stories: int = 5) -> str:
10    """Fetch and return the top stories from HackerNews.
11
12    Args:
13        num_stories: Number of top stories to retrieve (default: 5)
14    Returns:
15        JSON string containing story details (title, url, score, etc.)
16    """
17    stories = [
18        {
19            k: v
20            for k, v in httpx.get(
21                f"https://hacker-news.firebaseio.com/v0/item/{id}.json"
22            )
23            .json()
24            .items()
25            if k != "kids"
26        }
27        for id in httpx.get(
28            "https://hacker-news.firebaseio.com/v0/topstories.json"
29        ).json()[:num_stories]
30    ]
31    return json.dumps(stories, indent=4)
32
33
34agent = Agent(
35    model=OpenAIResponses(id="gpt-5.2"),
36    dependencies={"top_hackernews_stories": get_top_hackernews_stories},
37    add_dependencies_to_context=True,
38    markdown=True,
39)
40
41agent.print_response(
42    "Summarize the top stories on HackerNews and identify any interesting trends.",
43    stream=True,
44)

Set up your virtual environment

1uv venv --python 3.12
2source .venv/bin/activate

1uv venv --python 3.12
2.venv\Scripts\activate

Install dependencies

1uv pip install -U kern-ai openai httpx

Export your OpenAI API key

1export OPENAI_API_KEY=your_openai_api_key_here

Run Agent

1python add_dependencies_to_context.py