Laya / System One

Laya is the open-source System One model served on SGL Grid as convaiinnovations/laya. Use it when you need structured decisions instead of chat text: route choices, scores, or compact noul answers.

Laya is not a chat-completions model. Call POST /v1/systemone, not POST /v1/chat/completions.

Quickstart

curl "https://grid.x402compute.cc/v1/models?type=systemone" \
  -H "X-API-Key: $SGL_API_KEY"

curl -X POST https://grid.x402compute.cc/v1/systemone \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $SGL_API_KEY" \
  -d '{
    "model": "convaiinnovations/laya",
    "state": {
      "ticket": "Enterprise customer cannot access billing exports",
      "customer_tier": "enterprise"
    },
    "questions": {
      "route": {
        "type": "choice",
        "instructions": "Choose the best team to handle this ticket.",
        "criteria": {
          "billing": "Billing, invoice, refund, or account credit issue.",
          "support": "Product defect or technical troubleshooting."
        }
      },
      "urgency": {
        "type": "score",
        "instructions": "Score urgency from 0 to 1.",
        "criteria": ["customer blocked", "money at risk", "security risk"]
      }
    }
  }'

Typical response:

{
  "object": "systemone.result",
  "model": "convaiinnovations/laya",
  "answers": {
    "route": { "type": "choice", "choice": "billing", "confidence": 0.83 },
    "urgency": { "type": "score", "score": 0.7 }
  },
  "usage": { "input_tokens": 178, "output_tokens": 0, "cost_usd": 0.000001 }
}

Privacy Modes

Plain POST /v1/systemone validates and routes plaintext through the orchestrator, then dispatches only to attested confidential nodes. This is private at the node/runtime layer. For end-to-end private Laya, use the reserved sealed flow:

POST /v1/systemone/reserve with model and input_tokens_upper_bound.
Encrypt locally to the returned node X25519 key.
Submit the sealed payload to POST /v1/systemone.
Decrypt the sealed result locally.

In the sealed flow, the orchestrator sees ciphertext while the node decrypts locally and runs Laya inside its serving boundary.

Limits

Model id: convaiinnovations/laya
Aliases: laya, laya-system-one, systemone-laya
Question types: choice, score, noul
Max state: 64 KiB
Max questions: 32
Max criteria per question: 64
Max input context: 16,384 tokens
Max request body: 256 KiB
Max response body: 1 MiB

Operators

Node operators can serve Laya from Singularity Node app v1.7.4 or newer: open Models → System One → Laya and connect. The app manages a local loopback sidecar at 127.0.0.1:8765 and starts the normal SGL node service with the System One runtime attached. See Node setup for the operator path.

Overview

User Guide

Slayer Agent

Core flows

Staking Engine

Singularity Cloud Network

Contract Addresses

ERC8004

Skills

Singularity Payments

Agentic Fundraisers

Enterprise

Quickstart

Privacy Modes

Limits

Operators