nvbn/mrw.wtf

This repository builds mrw.wtf, a service that helps people find the perfect reaction GIF. You know the feeling, someone says something online and you want to reply with a GIF that captures your reaction. Instead of scrolling endlessly through GIF libraries, this project lets you search with natural language and get matched to a relevant clip. The project is actually a collection of smaller services working together. One piece scans Reddit and Imgur to gather GIFs. Another analyzes the text you type so the system understands what kind of reaction you're looking for. A public API ties everything together, and there's a web frontend plus mobile apps for Android and iOS. There's also a Telegram bot, so you can find and share GIFs directly in chat. The search itself relies on Elasticsearch, which is a specialized tool for quickly finding relevant results from large collections of data. The main audience is people who frequently communicate with GIFs, anyone who hangs out on Reddit, Telegram, or similar platforms and wants a faster way to find the right reaction. "MRW" stands for "my reaction when," a common phrase people use when posting reaction GIFs, which is where the project gets its name. The README doesn't go into much detail beyond listing the components and mentioning Elasticsearch. There's no setup guide or usage instructions included, so getting it running would require some familiarity with the underlying technologies. The choice to split everything into separate services suggests the project was designed to be modular, but the README doesn't explain the reasoning behind that architectural decision.