Ad Agent

The Idea & The Problem

It started with a need to build a simple tool for myself and friends to find new listings faster. The main requirement was speed and stealth (to avoid bans). Existing RSS solutions only updated once an hour, which was useless for catching hot deals. I couldn't find any free open-source tools that met my criteria either.

Market Analysis & Architecture

I conducted an in-depth analysis of multiple leading classifieds platforms. Most relied on legacy HTML rendering, which proved inefficient for high-frequency data processing.

To overcome this, I architected a custom data aggregation layer. By identifying and interfacing directly with internal data streams, I unified dispararate sources into a clean, structured format. This architectural shift allowed for sub-second data retrieval, bypassing the overhead of traditional web scraping.

Platform Integration

I chose Telegram for its instant notifications and seamless integration on iOS and Android. Initially, it was a command-based bot, but that proved too difficult for average users. So, I completely overhauled the concept and implemented a graphical interface (Mini App) directly within the chat.

Now users get a convenient UI to set up filters, while the bot handles the monitoring and alerting.

Current Status & Challenges

The Frontend is fully ready and deployed. The current challenge is optimizing the Backend to handle high loads. The system needs to process up to 200 requests per second to effectively filter out paid and "top" listings, ensuring only relevant new content reaches the user.