Platform experiment workbench

Test the primitive.
Keep the evidence.

Compare unfamiliar AI and onchain platforms without turning enthusiasm into a production decision. Record what happened, expose the gaps, and leave with the next test.

Local and keyless. Never enter private keys, seed phrases, wallet credentials, or API secrets. Scores are planning aids, not security or investment advice.

Experiment setup

Compare the evidence

Platform APlatform B

Experiment goal

Integration effort Documentation clarity Reproducibility Security confidence Breakage severity

Observed evidence Breakage or unresolved risk

Open scoring model

Evidence before ecosystem gravity.

Readiness= docs 15% + reproducibility 30% + security 30% + effort 10% + evidence 15% - breakage penalty

A severe breakage caps the recommendation until it is reproduced and reviewed. Text is stored locally and never transmitted by this page.