That is the idea of my recent failed project. Yet another project that going to blow your mind in... never.

As a human, I sometimes find interesting and secondhand stuff or something rare. But I can't check the web search every day, so I hope there is a tool to help me search for that item every day and notify me when it's found.

Sounds great! The tool will run in the background, diligently looking for your item on the web day by day. And one day, the system finds it! Ding ding! You receive a notification to look for the item.

I put a lot of effort into building this system and hope it becomes a valid SaaS. However, there are some obstacles that I don't want to overcome. Let me rephrase that: "don't want" instead of "can't". And that's the story I'm sharing today. It won't be a business insight story, but rather a pure technical story.

The keywords for this system are:

• Clean architecture

• Golang

• Typescript

• Postgres

• Queue table

• Monolith

• Mono repo

• Chromedp

• Proxy pool

• NLP

• BERT

• Vector database

• Token bucket rate limiter with Postgres only

• React

• React-query

This system also comes with some challenges in crawling data problems and NLP problems.

If anyone is interested in this topic, I will continue to write more until the end of this series.