The idea of Wienerdog was born out of my experience setting up my own simple but effective system of memory, self-improving skills and hooks for Claude Code and Codex. As I was teaching my friends and colleagues how to set up their own I found myself automating more and more of my system setup and finally I decided to publish it on GitHub to help others get more out of their AI usage.

So what is Wienerdog? Simply put, it is just files — no daemon, no server, no telemetry: a collection of instructions and skills that you can install by pasting just one line (npx wienerdog@latest init) and it will give your Claude Code / Codex:

- a proper base information CLAUDE.md / AGENTS.md file created by doing a simple, friendly interview with you (existing files are respected — Wienerdog only writes inside its own clearly-marked block). - a persistent, markdown-based memory vault following the PARA convention, similar to the Obsidian vaults many folks use. Already have a vault? No problem, you can just point it to it. - automated ‘hooks’ that will pull in key information (who you are, what you do, etc.) into each session and detect and save critical new information into your persistent memory. - automated daily ‘dreaming’ runs that digest the previous day’s sessions and store their key information into memory. If your computer is off or asleep at that time, the dreaming run will catch up when you're back. - repeating task patterns are identified and turned into reusable skills which can be further improved automatically. - if you are using both Claude Code and Codex (like I do) they will be able to read from and contribute to the same memory vault. You are not locked in to any vendor. - a basic (and optional) integration with the Google suite to give your AI access to your Gmail, Calendar and Drive — read-only and draft-only by default.

Everything Wienerdog does can run on your existing Anthropic / OpenAI subscription, in line with their terms. The project is open-source on GitHub and free forever under the MIT license. I do this for fun and to give a little something back.

I have put quite a bit of work into Wienerdog already and am happily using it myself for its own development (dogfooding — get it?). In fact, most of Wienerdog's code was written by AI agents following its own spec system, using the memory vault that lives in the repo. With that said, it is still early and a work in progress that could use more eyes. I'd love for you to try it out and ask questions, request new things to add or even contribute on GitHub if you have the time and capacity.

P.S. Why the name Wienerdog? Because I love these little tube-shaped snouty sausage clowns :)