Limited Promo: 1,000 Free Calls/mo on RapidAPI →
Created by Harshad Jadav • Convai Laya 322M

System 1 AI Decisions.
Zero Generative Latency.

Stop waiting 2,000ms for slow, hallucinating generative LLMs to make structured decisions. Laya evaluates categorical choices, urgency scoring, and prompt guardrails in a single mathematical pass in under 35 milliseconds.

< 35ms
Average Latency
0%
Hallucinations
100+
Languages (mmBERT)
Zero
Data Retention

Capabilities

Engineered for Workflow Automation

POST /v1/triage

Support Ticket & Email Triage

Classifies customer inquiries into departments (billing, technical, account), grades urgency level (0-2), and flags churn/cancellation risks.

POST /v1/guard

AI Agent Prompt Injection Guard

A sub-30ms security perimeter that detects jailbreaks, command overrides, and role-play attacks before queries reach expensive generative LLMs.

POST /v1/filter/spam

Spam & Phishing Filter

Real-time classification distinguishing legitimate inquiries from commercial marketing junk, phishing scams, and credential theft attempts.

POST /v1/sentiment

Sentiment & Frustration Analyzer

Evaluates text sentiment along with customer frustration and anger intensity score to trigger priority routing for escalated inquiries.

POST /v1/moderate

Content Moderation Gate

Zero-shot multi-label safety detection flagging toxic insults, hate speech, and explicit content before storing user-generated content.

POST /v1/decide

Custom Typed Decision Engine

Universal endpoint allowing arbitrary choice, score, or boolean questions on any text or state payload in a single forward pass.

Quickstart

Call via RapidAPI in Seconds

curl-example.sh
POST /v1/triage
curl --request POST \
  --url https://laya-ai-api-gateway-lightning-decision-engine-api.p.rapidapi.com/v1/triage \
  --header 'x-rapidapi-host: laya-ai-api-gateway-lightning-decision-engine-api.p.rapidapi.com' \
  --header 'x-rapidapi-key: YOUR_RAPIDAPI_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "subject": "Billing issue on invoice #1024",
    "body": "I was double charged this morning. Please refund immediately or cancel my plan."
  }'
// Sample Output (< 35ms):
{
  "department": "billing",
  "urgency_level": "critical",
  "is_churn_risk": true,
  "recommended_priority": "critical",
  "latency_ms": 28.4
}

Attribution & Architecture

Creator & Foundational Model

Harshad Jadav
Gateway Creator & Architect

Designed, built, and deployed the production REST API runtime, marketplace security middleware, and cloud hosting infrastructure on Red Hat OpenShift.

Convai Innovations & BFC
Model Creator & Research

Powered by the open-source convaiinnovations/laya foundation model — a 322M parameter multilingual bidirectional encoder (mmBERT) tuned for calibrated categorical decisions without generative overhead.

FAQ

Frequently Asked Questions

Answers to common questions regarding latency, architecture, privacy, and integrations.

What is Laya AI System 1 Decision Gateway?
Laya AI System 1 Decision Gateway is a deterministic, sub-35ms decision engine built on the 322M parameter mmBERT encoder that evaluates categorical choices, urgency levels, and security guardrails in a single mathematical forward pass without generative LLM hallucination or latency.
How fast is Laya Decision Gateway compared to traditional generative LLMs?
Traditional generative LLMs take 1,500ms to 3,000ms to generate and parse structured JSON. Laya completes decision inference in under 35 milliseconds, making it over 50x faster and ideal for high-throughput automation workflows in n8n, Make, and backend microservices.
Does Laya retain or train on customer prompt data?
No. The gateway enforces a strict zero-retention policy. All customer payloads are processed strictly in ephemeral RAM and immediately deallocated upon response completion. No customer data is ever saved to disk or used for training.
Is there a free tier to test the API?
Yes, developers can subscribe to the free tier on RapidAPI which includes 1,000 free monthly requests during the introductory period.
Can I integrate Laya with n8n, Make.com, or custom microservices?
Yes. The API uses standard REST JSON interfaces over HTTPS. You can invoke any endpoint using native HTTP Request nodes in n8n or Make.com, or standard HTTP client libraries in Python, Node.js, Go, or cURL.