Laya AI · System One

Laya AI — System One Decision Model

Typed decisions for agents and software — choice, score, and yes/no in one forward pass. Hosted prepaid API here; Apache 2.0 weights if you run locally.

Try playgroundGet API keyIncludes 10,000 tokens each monthPricingDownload

Typed decisions, live

Pick a scenario, edit the input, run the real API. Choice, score, and yes/no come back with confidence.

Anonymous free runs: 1. Then sign in with Google or paste a laya_ key.

Browse all tools →

POST /v1/support/triage Live

State — the input software gives Laya

190c

Questions — typed decisions you get back

choice department

Which queue should handle this?

score urgency

How urgent is this request?

noul churn_risk

Does the user threaten to cancel or leave?

noul refund_requested

Does the user explicitly request a refund?

Sample answers

sample · not yet run

sample · not yet run

choice department

billing72%
technical9%
account9%
other9%

score urgency

not urgent12%
soon12%
today12%
critical64%

noul churn_risk

yes81%
no19%

noul refund_requested

yes81%
no19%
laya_Key after Google sign-in
No GPU opsThe server holds the inference key
≈ $0.0004per 1,000 input tokens
10k_monthly10k_monthly · 10,000 input tokens / month

Laya vs Jev

0.766 belongs to the fine-tuned laya-typed-decisions checkpoint. The English and multilingual bases score 0.362 and 0.352 on those same 2,000 decisions, under a 0.461 majority baseline. pip install does not apply that fine-tune.

README table “Laya (with routing) vs Jev”. Jev figures are published third-party numbers, not remeasured here.
Jev 1.13.0Laya (routed)
typed-decisions, 2,0000.7270.766
AG News, 4 labels0.9100.950
DAIR Emotion, 6 labels0.4800.595
Banking77 (72 vs 77)0.8700.425
ECE, lower better0.2460.081
p50, 1 question236–276 ms32.8 ms

Full benchmarks

Run locally

python -m pip install laya

Get Started · NandhaKishorM/laya

Laya AI — System One Decision Model