These past couple of days, Vibe Coding made a useless little tool and casually named it slug2text. Here’s the link: https://slug2text.pages.dev The functionality is super simple: it helps you compress large chunks of text into a single link without losing any data. Just save the link, and when you open it again, the original text will automatically decompress and restore itself. The compression rate depends on the size and redundancy of the text, but based on my tests, it can handle something as long as a thesis. However, due to URL length limitations, you can’t fit in too much content. Honestly, the main reason for making this was out of two curiosities: 1. How extreme can GPT-5.2-Pro’s compression solutions get? 2. How good is GPT-5.2’s programming ability in the Codex environment? Conclusion: I can’t really evaluate the compression rate, but based on its own assessment, it seems pretty decent. As for the engineering implementation, there were no issues—Codex passed everything on the first compile.