Digestify

Digestify is a Python library that generates comprehensive news digests and content summaries for specific topics using OpenAI and the WebQuest MCP server. It intelligently gathers latest updates, including processing video transcripts, to provide structured summaries.

Installation

Installing using pip:

pip install digestify

Installing using uv:

uv add digestify

Configuration

Digestify requires the following environment variables to be set (either in your environment or a .env file):

• OPENAI_API_KEY: Your OpenAI API key.
• WEBQUEST_MCP_ACCESS_TOKEN: Access token for the WebQuest MCP server.
• WEBQUEST_MCP_URL: URL for the WebQuest MCP server.
• OPENAI_MODEL: (Optional) The OpenAI model to use (default: "gpt-5.2").
• MAX_TOOL_CALLS: (Optional) Maximum number of tool calls allowed (default: 5).
• MAX_OUTPUT_TOKENS: (Optional) Maximum number of tokens for the output (default: 100000).

Usage

Here is a simple example of how to use Digestify to get a digest for a topic:

import asyncio
from digestify import Digestify, Topic

async def main():
    # Define the topic you want a digest for
    topic = Topic(
        name="Artificial Intelligence",
        description="Latest advancements in AI, machine learning, and LLMs."
    )

    # Initialize the client (automatically loads settings from env/files)
    client = Digestify()

    # Get the digest
    digest = await client.get_stories(topic)

    # Print stories
    for story in digest.stories:
        print(f"Title: {story.title}")
        print(f"Published: {story.published_at}")
        print(f"Summary: {story.content}")
        print("References:")
        for url in story.reference_urls:
            print(f"- {url}")
        print("-" * 20)

if __name__ == "__main__":
    asyncio.run(main())