Skip to content

Automatically collect and summarize articles from WeChat Official Accounts, using LLM and Feishu doc.

Notifications You must be signed in to change notification settings

DayDreammy/recruiment_wechat_airticle_summary

Repository files navigation

Pandora WeChat Article Collector and Summarizer

Automatically collect and summarize articles from WeChat Official Accounts.

Features

  • Batch collection of WeChat article links using Quicker
  • Article downloading using wechatDownload
  • Article summarization using Large Language Models (LLM)
  • Automatic upload of summaries to Feishu (Lark) documents and Airtable

How It Works

1. Collecting Article Links with Quicker

We use Quicker to simulate mouse and keyboard operations for batch collection of article links.

Link Collection Demo

Output:

Links.txt

Note: Currently optimized for 2560×1440 screens. Other resolutions may require adjustments.

2. Downloading Articles

We use the open-source project wechatDownload to download the articles.

3. Summarizing Articles

Articles are summarized using Large Language Models.

4. Uploading Summaries

Summaries are automatically uploaded to Feishu (Lark) documents and Airtable for easy access and organization.

Complete Workflow

Full Quicker action: Automatic WeChat Article Collection and Summarization

Full Workflow Demo

Output:

Article Files

Set this to run daily for automated WeChat Official Account monitoring and collection (RSS-like functionality).

Use Case

Pandora Job Information Sharing

As of 2024-03-28:

Job Information Dashboard

Future Improvements

  • Refactor and optimize code
  • Improve Quicker action to be more adaptable to different screen resolutions
  • Enhance PDF summary operations and database integration
  • Integrate with larger RSS projects
  • Expand to more diverse information sources (news sites, podcasts, blogs, Bilibili videos, etc.)

Vision

Our ultimate goal is to facilitate smoother information flow and provide an AI assistant to help process information overload:

  1. Summarize content from various sources to create a personalized RSS-like experience
  2. Allow users to stay informed while spending less time on information consumption

Notes

  • This project demonstrates that sometimes the most effective anti-crawling measures are the simplest, human-like approaches
  • The quality of input data significantly shapes the model's (and human's) potential
  • We aim to make information more accessible and manageable in our increasingly data-rich world

About

Automatically collect and summarize articles from WeChat Official Accounts, using LLM and Feishu doc.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages