Hello, HN! If you've been trying to integrate LLMs into WhatsApp, you've probably noticed that the massive blocks of text that the AI often responds to, besides being in Markdown instead of the format WhatsApp supports, don't provide a good user experience.
I just published an open-source repository that solves this problem:
- Converts Markdown to WhatsApp's format - Intelligently splits long texts into shorter fragments (without cutting lists, links, emails, and with syntax understanding) - Supports tables, so they look good - Runs 100% locally in your project, doesn't use external libraries, and has 100% test coverage
You can start using it with a simple 'npm install'.
If you find it useful, I'd appreciate it if you gave the repo a star :