Tailcat: Tailscale's Network Debugging Tool

Tailcat is an open-source network debugging tool from the Tailscale team. It works just like classic netcat, but runs over Tailscale's data plane without needing the control plane.

For ops and network engineers, this means you can test TCP/UDP connections, verify port forwarding, and troubleshoot reachability issues directly within a Tailscale network — no need to spin up a full Tailscale client. It's a single Go binary that you can drop onto any machine and run.

The community has been asking for this: previously, doing similar operations inside a Tailscale network required workarounds. Now there's an official native solution.

GitHub: https://github.com/tailscale/tailcat

ThreeUI: An Open-Source Three.js 3D Component Catalog

ThreeUI is a Three.js 3D UI component collection that hit 4000+ GitHub stars shortly after launch. Unlike typical component libraries, every component comes with a full interactive online demo — you can drag, rotate, and tweak parameters in the browser before deciding whether to use it in your project.

Components cover React, WebGL Shaders, and various tech stacks. If you're building a front-end project that needs 3D visual effects (product showcases, data visualization, immersive pages), ThreeUI's components can be copied directly into your project, saving you the effort of writing Three.js rendering logic from scratch.

The site itself is beautifully designed, showing how 3D components look in real-world scenarios.

Website: https://threeui.com

acceptmarkdown.com: Clean Markdown for AI Agents

AI agents scraping web content has always been a pain point — HTML parsing produces too much noise, and LLMs waste tokens processing irrelevant navigation bars, ads, and script code.

acceptmarkdown.com proposes a straightforward solution: web pages negotiate content via the HTTP Accept: text/markdown header. Browsers receive HTML as usual, while AI agents automatically get clean Markdown. No need for a separate API, no extra parsing on the client side.

This approach follows the same philosophy as robots.txt and sitemap.xml — a minimal protocol convention that solves a widespread problem. It's still early, but the direction is right: as AI agents increasingly consume web content, a native Markdown distribution mechanism will become more important.

Website: https://acceptmarkdown.com