Free browser tool
Content API Cutoff Check
Check a representative code sample for common Google Content API for Shopping migration signals. The deterministic scan runs in your browser; pasted code is not uploaded or stored.
Run the free scanAI venture lab / autonomous experiments
An AI venture lab where experiments learn to evolve.
MicroForge uses AI agents to build prototypes, test them with real signals, and feed what survives back into the next cycle.
Some experiments become games. Some become tools. Some stay private. Some disappear back into the forge.
ExperimentalNo accountsNo adsPrivacy-light signals
MicroForge loop
Public-safe experiment engine
1
Generate
agents explore ideas
2
Build
small working artifacts
3
Signal
play and feedback
4
Evolve
mutate or retire
Not everything is public.
Staged experiments stay private until they are safe, working, and approved for exposure.
1. Current Public Experiments
A public window into the forge.
Public lane
Forge Arcade
The first public lane: small browser games that can be played instantly and judged honestly.
Enter ArcadeExperimental
Crayon Traffic
Route crayon-drawn cars through a busy intersection. Current human-signal game winner.
Play nowRunner-up
Open
Published only
Product Experiments
Browse the approved tools and business experiments that have crossed MicroForge's public-safety boundary.
View index2. How MicroForge Works
A simple loop with hard feedback.
01
Ideas enter the forge
Agents explore small bets, weird angles, games, tools, and product shapes.
02
Prototypes get built
The best candidates become working artifacts, not pitch-deck fiction.
03
Signals push back
Play, feedback, and usage expose what internal reviews miss.
04
Winners evolve
Strong experiments mutate. Weak ones are retired without ceremony.
3. Public / Private Boundary
What you see here is only the safe surface.
The operator forge stays private.
Internal dashboards, staged prototypes, run logs, agents, and review queues are not part of the public website. Public pages only link to approved routes.
Signal-driven
The system learns from use, not vanity numbers or internal excitement.
Small enough to judge
Experiments start compact so people can understand them quickly.
Public/private boundary
The public site shows approved experiments. The operator forge stays private.
Automation with brakes
Agents can build and critique, but external actions remain governed.
4. Lab Notes