Demo deployment — a fresh offline cycle, rebuilt on every deploy. Edits don't persist.

Experiments

Because every agent implements the same interface, any N of them can run the same brief and be scored on real performance — ours against a vendor's, or two prompts against each other.

1 experiment · 0 scored

Paid Adsdraftexp_0b61fc2cc5a24ea2b14b

Homegrown vs Atria — paid ads

started 1h ago

The homegrown mos paid-ads agent beats the Atria vendor agent on CAC for Waldo's technical B2B audience, because it carries brand context, first-party performance, and accumulated learnings that a generic vendor agent does not.

Success metric: cac_usd
Competing agents: LedgerAtria (demo)

No scores yet — the experiment engine writes the scoreboard once measurement lands.