Hi Product Hunt! 👋 If you build with LLMs, you've lived this pain: You're debugging a complex prompt, so you log it. You open your console or cloud logs and... it's one giant, ugly line. `\n` instead of actual line breaks. JSON payloads crammed inline. The system, user, and assistant turns all fused together into an unreadable wall of text. To figure out what you actually sent to the model, you end up manually un-escaping a string in a text editor. I got tired of doing that by hand, so I built FormatTools. Paste that logged mess on the left — get a clean, color-coded conversation on the right: • Unescapes `\n`, `\t`, and quotes back into real formatting • Pretty-prints inline JSON payloads automatically • Splits system / user / assistant / tool messages into separate, labeled blocks • Works with OpenAI, Claude, and other popular LLM API formats Security-wise, it runs 100% client-side in your browser. No signup, no backend — your prompts never leave your machine, which is crucial when you're pasting real production logs. It's completely free. This started as my own internal debugging tool, so I'd genuinely love to hear from other LLM devs: what features would make this your daily driver? A prompt diff view? Token counts? Let me know in the comments! 👇 (P.S. There are also Markdown and HTML formatters on the site if you need them — but the prompt formatter is the one I actually built this for!)

FormatTools

Make your logged LLM prompts actually readable

Make your logged LLM prompts actually readable