System 1. System 2. System null.
shrug.ai: stupid, fast
You’ve heard of System 1 (fast) and System 2 (slow) thinking. shrug.ai is System null: no thinking at all — and, conveniently, your null hypothesis. It answers decision and judgment calls with pure randomness: no model, no reasoning, no latency. Think of it as placekitten for decision responses, a drop-in placeholder that hands back a perfectly-shaped answer without thinking about it for a second.
Free up to 1,000 requests/month. No model, no latency, no excuses.
Two ways to shrug
A mock that speaks the protocol
Swap shrug.ai in wherever you'd call a real decisions or judgment endpoint — in tests, local dev, or a demo. Same request and response shape, instant and free, no reasoning-model keys to manage and nothing to go flaky. It's the server mock that already talks like the real API.
A baseline for the real thing
Point your eval harness at shrug.ai to get the chance-level floor. If OpenAI Decisions, TypeSafe System One, or your own System 1 model can't beat a random shrug, that's a result worth knowing — and now you can measure it.
How it works
Drop in a baseline
Point your OpenAI Decisions or TypeSafe System One client at shrug.ai. Same request shape, same response shape — no code changes beyond the base URL and key.
Get a principled random draw
Every question is answered by sampling, not thinking: choices, predicates, and scores come back with calibrated probabilities. That is your chance-level floor.
Bias the dice (optional)
Add a shrug.distributions block to weight options by name or set p_true for a predicate. Normalized automatically, so you control the null hypothesis.