Global Software AI
The Lab

Where ideas go to mutate.

Part mad-science bench, part AI playground. This is where Global Software AI tinkers — prototypes, experiments, and apps that exist mostly because we wondered "what if?"

Warning: unstable prototypes ahead. Ideas may mutate without notice.

Experiment log

Current experiments

Stable-ish

AI security learning game

Threat Quest: Defend Byteburg

Explore the OWASP AI threat model, discover risks, choose mitigating controls, and defeat surprise Auditor Monsters by producing the right evidence.

Play experiment
Actively bubbling

AI security arcade game

Threat Tree Arcade: Neon Breach

Learn the same OWASP AI threat-model concepts through a faster, animated arcade challenge with less reading and more action.

Play experiment
Actively bubbling

Generative audio & visuals

AImpressions

Feed it a text or chat history and it distills the vibe into an original song — with matching art, a cartoon, or a full video. Your group chat, but make it a music video.

Stable-ish

Hospitality ops

Bar management suite

Inventory, cash management, and operations tools for bars — born behind a real bar and tested against real Friday nights.

Actively bubbling

RegTech tinkering

Compliance mini-tools

Small utilities poking at regulated-industry workflows. The promising ones feed their findings back into CompylotAI.

Stable-ish

Legal automation

Founder Agreement Builder

A 13-step wizard that interviews founders and generates a founder agreement as a document — equity, vesting, IP, and the awkward what-if-someone-leaves parts.

Play experiment

No production servers were harmed in the making of these experiments.

Field notes

Experiments get written up on the blog — what worked, what exploded, and what we learned along the way.

Read the lab notes

Got a weird idea worth testing?

The best experiments start as someone saying "this is probably silly, but…". Tell us yours.

Ping the lab