How to serve Qwen3.6-35B-A3B, a Mixture-of-Experts model with 3B active parameters, on an RTX 5070 Ti using llama.cpp. Full config, performance numbers, and the flags that make it fit.
Running a 35B MoE Model on a 16GB Consumer GPU


How to serve Qwen3.6-35B-A3B, a Mixture-of-Experts model with 3B active parameters, on an RTX 5070 Ti using llama.cpp. Full config, performance numbers, and the flags that make it fit.

A thread on the Fediverse infected my curiosity with the question: How small of a server can GoToSocial run on and still give a decent user experience for a single-user instance? Let’s find out.

A step-by-step guide to self-hosting the Gotify notification service using Docker and PostgreSQL.

A guide for self-hosters to deploy Uptime Kuma using Docker and Docker Compose.

Guide to deploying ActivePieces with Docker Compose, including customizations and future improvements.

Notes on Bookstack


In this guide, we’ll walk through the process of setting up Huginn, a powerful open-source system for automating tasks and managing workflows, using Docker. We’ll use MariaDB as the backend database and go over how to configure it with simple environment variables.
Before we start, you’ll need:
If you’re on macOS, you will need to install the uuid command line tool through Homebrew. You can do so with the following command:

notes on n8n - a free workflow automation tool

notes on Traefik, an open source reverse proxy and ingress controller
Computers are getting smaller, faster, and cheaper. Entire homelabs can now fit inside of a 10 inch wide rack.