These days, artificial intelligence or AI is not just a technological concept—it is steadily evolving into our personal assistant, research companion, and information analyst. In particular, the concept of the “AI agent” is becoming increasingly popular. An AI agent refers to an automated system that can collect information, analyze it, make decisions, and perform specific tasks—often with minimal human intervention.
I have also put this idea into practice by creating a small yet effective AI agent whose main task is to find relevant articles or writings for me. As a columnist and op-ed writer, I need to know every day—which topics are being discussed around the world or especially in Bengali, which subjects are at the center of attention, and what perspectives other writers are presenting. However, the reality is that spending hours each day browsing various websites to collect this information is extremely time-consuming and exhausting.
As a solution to this problem, I have developed an AI agent that automatically collects new articles from various reliable sources on the internet. Here, using the Python language, necessary information is gathered from web pages via the requests and BeautifulSoup (bs4) libraries. For websites that rely on JavaScript and do not provide data directly, Playwright is used to load pages in a virtual browser and extract the required information.
But simply collecting information is not enough. The most important aspect is relevance. I don’t want to see every type of news or article; I only want those topics that are related to my interests and assist in my writing. This is where AI comes in. My system features a classification mechanism, where each article is checked using an AI model to determine how closely it relates to my designated topics. As a result, irrelevant or less important content is automatically filtered out.
Another critical feature is avoiding repetition. To prevent the same article from appearing multiple times, a list is maintained to store previously seen URLs. This ensures that each time, I only receive new and unseen content.
Finally, the results of this entire process are delivered to me in a very simple way—via Telegram. Using the Telegram Bot API, my AI agent organizes the new articles in a specific format and sends them directly to my mobile Telegram channel. No matter where I am, I can instantly find out what new content has been published and which topics are currently being discussed.
The whole system is set to run automatically at fixed intervals using Linux cron, so I do not have to operate anything manually. It works on its own—like a silent assistant.
Simply put, this AI agent acts as a “digital research assistant” for me, constantly scouring the internet for my preferred topics, analyzing them, and notifying me accordingly. As a result, instead of getting lost in a flood of unnecessary information, I can focus directly on important and relevant content.
In today’s world, where the volume of information is increasing constantly, such an AI agent not only saves time—it also helps sharpen our thinking. Especially for those involved in writing, research, or analytical work, this can become a unique and invaluable technological aid.
See the agent’s full code on GitHub: https://github.com/drmashiur/oped-agent/
Leave a comment