Blog

Notes from the workshop.

Where documentation, agents, and the record meet. New posts land here first.

August 11, 2026Mari is open sourceThe product knowledge platform we've been building is now Apache 2.0. Self-hosted, a dozen connectors, cited answers, and facts with owners. Here's what you get and how to run it.Eric Disque · 2 min read
August 11, 2026Zero to knowledge base in 10 minutesThe honest getting-started walkthrough: clone, compose up, redeem the setup token, connect GitHub, and watch the agent answer with its work shown. Real commands, real screens, measured timings.Eric Disque · 4 min read
August 3, 2026Why we built MariThe knowledge was never missing. The accountability was. Why we built a product knowledge platform where every fact has a source, an owner, and a freshness state.Eric Disque · 4 min read
May 8, 20265-layer agent runtime - three document layersMemory, knowledge, and delegation are the three Claude Code layers most teams quietly lose control of. They are documents in disguise — and they rot the same way every other doc rots.Daniel Henneberger · 8 min read
May 8, 2026Skills are documents tooLLM knowledge is jagged. Treating skills as composable, validated documents is how we make the edges less surprising.Daniel Henneberger · 4 min read
May 8, 2026Your docs are already being read by AINearly half of documentation traffic now comes from AI agents, not humans. Most docs were not written for that audience. Here is what that actually breaks.Eric Disque · 5 min read
May 7, 2026The state of docs platforms (and where Mari fits)Mintlify, Fern, ReadMe — and why none of them write the docs for you. A frank, opinionated tour.Daniel Henneberger · 4 min read
May 7, 2026We need background agentsThe next leverage point for engineering teams isn't a smarter chat window. It's a fleet of patient, scoped agents working overnight.Daniel Henneberger · 2 min read