Building an LLM Agent That Writes a Daily Social Feed Digest

Step by step

1. 1

   Define a fetch tool

   Expose a single function such as fetch_page(url) that returns markdown. Agents use simple, well-described tools more reliably than raw HTML.

2. 2

   Fetch public pages only

   Point the tool at publicly available profile or list pages, and respect each platform's terms of service and robots.txt.

3. 3

   Summarise in a fixed format

   Ask the model for a digest with sections such as top themes, notable posts with links and open questions, so day-to-day output stays comparable.

4. 4

   Schedule the run

   Run the agent once a day from a scheduler and store the digest, so the model never needs to remember previous days.

Example

import requests

def fetch_page(url: str) -> str:
    """Tool for the agent: return a public page as markdown."""
    response = requests.post(
        "https://alterlab.io/api/v1/scrape",
        headers={"X-API-Key": "YOUR_API_KEY", "Content-Type": "application/json"},
        json={"url": url, "render_js": "auto", "formats": ["markdown"]},
    )
    return response.json().get("content", {}).get("markdown", "")

Frequently Asked Questions