Unofficial community documentation
Hosted Laya API — prepaid decide calls
Skip torch, T4s, and Router preload. Register, buy a pack, create a laya_ key, call /v1/decide and the tool templates. Typical call ≈$0.0004.
Priced per decision, not per chat token
System One returns choice / score / yes-no in one forward pass. A typical /v1/decide is about $0.0004 — a few tenths of a mill. Routing the same triage or label job through GPT/Claude chat completions usually costs several× that per call (prompt + completion), before you parse free text.
- Instant key — register, top up, mint a
laya_key in minutes. - No GPU ops — no torch install, no Hugging Face download, no Router preload.
- Tools included — triage, jailbreak gate, scam spot, and more on the same prepaid balance.
Want local weights instead? See Download and Get started. Benchmark tables: Laya vs Jev.
Per decide
≈$0.0004
typical /v1/decide · vs multi-cent chat-LLM triage
$5 starter
≈12.5k decides
enough to wire triage / gates in prod
Output
Free
Structured answers, not chat completions
Starter
Start$5
≈ 12,500 typical /decide calls
11,900,000 input tokens · 60 RPM
$10
$10
≈ 25,000 typical /decide calls
23,800,000 input tokens · 60 RPM
$20
$20
≈ 50,000 typical /decide calls
47,600,000 input tokens · 90 RPM
Staging
$50
≈ 150,105 typical /decide calls
142,900,000 input tokens · 120 RPM
Popular
Popular$100
≈ 420,168 typical /decide calls
400,000,000 input tokens · 120 RPM
Scale
$250
≈ 1,193,697 typical /decide calls
1,136,400,000 input tokens · 180 RPM
Volume
$500
≈ 2,626,050 typical /decide calls
2,500,000,000 input tokens · 240 RPM
Checkout
Register or sign in first, then buy a pack. Credits land on your Account balance; create a laya_ key there. Anonymous visitors get 1 free tool run; registered accounts get a small monthly starter grant — enough to verify, not to run prod.
or before checkout.
API base: https://api.layaaimodel.com. Manage keys on Account.
Unofficial community host — not a Convai / TypeSafe / Impossibl SLA. See Terms.