Pre-spend quote

Before you open a session: the live per-job fee, how many workers are eligible to serve the model RIGHT NOW (redundancy depth), the measured completion rate + latency, and when your fee auto-refunds if a worker stalls - one decision object. Read-only, no wallet, Mainnet.

Request

ModelAny whitelisted model on the selected network.

Read-only · Mainnet

Quote

Pick a model and Get quote to see whether it's worth routing to, before spending anything.

The SDK behind it

TypeScript

import { LightNode } from "lightnode-sdk";

const ln = new LightNode("mainnet");
const q = await ln.preInferenceQuote("llama3-8b");

if (!q.routable) throw new Error(q.verdict);
// fee, redundancy depth, reliability, latency, refund timing - before you spend
console.log(q.feeLcai, q.eligibleWorkers, q.completionRate, q.p95, q.refundWindowSec);

ln.preInferenceQuote(tag) - route only when q.routable is true.