A language-learning app in production.
10Languages
2,900+Automated tests
EUHosting · GDPR-compliant · payment live
Consulting · Munich
The studio.
The human.
The AI stacks.
Development, design, operations, marketing, support — carried by AI agents, steered by the human who signs off. What comes out of it runs in production.
The human
decides
AI agents
do the work
In production
what is built here
How it works
Everything that runs here runs on AI.
The human decides. The AI stacks do the work: one plans, one builds, one designs, one operates. The whole thing is steered through a chat channel — a phone is enough.
The human
Has the idea, decides, signs it off. The only point in the system where anything becomes binding.
The planning stack
An AI operator takes the idea in, plans it through together with the human, cuts it into individual assignments, checks the results against the rules and puts them up for approval.
The build stack
Several AI lanes build in parallel, each in a sealed-off workspace with its own audit record. Code, content, infrastructure — from assignment to approval.
The design stack
Interfaces take shape in the AI canvas: layout, typography and colour system are drafted in dialogue and iterated until they sit right. This page was built that way.
The operations stack
OpenClaw: agents with fixed business roles in continuous operation — management, DevOps, support, legal, finance, sales, marketing, content. They watch each other and report when something runs outside its limits.
All stacks share one memory and are dispatched by messenger (WhatsApp, Telegram) — assignment out, result back, from anywhere. Approvals stay tiered: what is reversible goes through, and the human decides at the business-critical gates.
From the studio
To make patterns in markets visible: when something recurs regularly, currently on Bitcoin and gold — and in which timeframes trading actually happens.
3layers
7hypothesis groups
BTC + gold
A decision instrument for day trading, not a piece of analysis writing: a surface where the picture across crypto assets, equity futures and the macroeconomic setting comes together, and which points out movements while they happen.
3asset classes
several sources per instrument
staleness gate
TradingSignals
LiveTo generate valid trade signals: indications you can lean on because they were examined before they were believed.
16hypotheses
5mechanism families
3running experiments
Infocluster
In progressThe knowledge layer AI agents work out of.
5components
separated per mandate
receipt per access
JANUS
In progressTo find out where the large market makers currently stand and what they are about to do — and whether anything in that could be acted on.
12houses
3jurisdictions
translate.bavarian.app
In progressA translator between German, English and Bavarian, in both directions.
3languages
key server-side
Agency craft, management in a media house, self-employed since 2008, running blockchain infrastructure first-hand since 2009. The track record explains why this studio works the way it does.
Read the track record →Contact
First contact goes by email or WhatsApp — no form and no stored data. Briefly describe what it is about; the reply comes from the human, not from an agent.