A single front door for all your AI agents and MCPs

The real learning comes from analyzing the incident while writing the document, not reading it; the document at the end is the residue of the learning.

AI tools have genuinely changed how fast teams can produce code, but they haven't changed how fast a codebase can safely absorb that code.

Benchmarking is hard. There are many ways to do it wrong and few to do it right.

But zooming out from any single system or harness, there are broad principles that should be applied to all benchmarking. Using these correctly makes it difficult to produce biased results.

Am I the world's best benchmarker? Certainly not. I invented the language balls, after all. But correctness and precision are important parts of PlanetScale's culture. We've spent considerable time learning the art of benchmarking, and are here to share best-practices.

Here, we're focusing primarily on benchmarking databases, but these principles apply to many domains.

When thinking about Service Level Objectives (SLOs) and contractual Service Level Agreements (SLAs) for availability, I always like to put the percentages into concrete numbers.

When working on code with GenAI assistance you need a better understanding of the system than when working without.

High-severity incidents expose how a team really works together, usually within the first ten minutes.

You’re in a 1:1 with your manager, and things are going just fine. You talk about the project and that other thing. Toward the end, she asks: “Anything else?”

And there is something else. You want to lead that new initiative. Or move to a different team. Or you’ve been thinking about what stands in the way of your promotion. The thought is right there, sitting in the back of your throat. You’re going to say it, and then… “Nope, all good.”

You get out of the call feeling a specific kind of regret. You rationalize it somehow and then tell yourself you’ll bring it up next time (you won’t).

mq is a command-line tool that processes Markdown using a syntax similar to jq.

It's written in Rust, allowing you to easily slice, filter, map, and transform structured data.

“In the age of AI, taste is the ultimate differentiator.”

Slumber is a TUI (terminal user interface) HTTP client. Define, execute, and share configurable HTTP requests.

MarkItDown is a lightweight Python utility for converting various files to Markdown for use with LLMs and related text analysis pipelines.

An infinite canvas for your code, terminals, browsers, docs, and AI agents.

Paneru is a MacOS window manager that arranges windows on an infinite strip, extending to the right. A core principle is that opening a new window will never cause existing windows to resize, maintaining your layout stability.

The open-source framework for AI SRE agents, and the training and evaluation environment they need to improve. Connect the 60+ tools you already run, define your own workflows, and investigate incidents on your own infrastructure.

Kubedock is a minimal implementation of the docker api that will orchestrate containers on a kubernetes cluster, rather than running containers locally. The main driver for this project is to run tests that require docker-containers inside a container, without the requirement of running docker-in-docker within resource heavy containers.

96 B200 GPUs. 12 nodes. A load balancer distributing requests across them.