Loading...
Loading...
Real mainnet data refreshed once a minute. Each card shows the SDK call that produced it.
Each card pairs a one-line SDK snippet with the data that call returns right now.
One call that summarizes the entire mainnet.
const ln = new LightNode("mainnet");
const stats = await ln.getNetworkStats();Per-worker completion + p50 latency over the last 500 jobs.
const workers = await ln.getWorkerStats(500, 5);| worker | jobs | complete | p50 |
|---|---|---|---|
| 0xA22C…0847 | 93 | 99% | 18s |
| 0x5CbA…078C | 78 | 100% | 18s |
| 0x86c7…9f85 | 50 | 92% | 18s |
| 0x1F89…5EB5 | 47 | 98% | 12s |
| 0x064c…6B00 | 47 | 89% | 24s |
Live whitelist with on-chain fee + max output tokens.
const models = await ln.getModels();| model | fee | max out | live |
|---|---|---|---|
| llama3-8b | 0.020 | 2,048 | yes |
| llama3-70b | 0.150 | 4,096 | yes |
Completion rate, p50 / p95 latency, incomplete count over the last 500 jobs.
const stats = await ln.getModelStats(500);| model | jobs | complete | p50 / p95 |
|---|---|---|---|
| llama3-8b | 500 | 94% | 18s / 36s |
Five stages of the inference path. Bridge / DAO / read-only have their own much shorter shapes.
SIWE handshake yields a JWT for the consumer gateway.
Worker selected. Session key wrapped with ECDH-P256.
Your wallet signs createSession on chain.
AES-GCM encrypted prompt uploaded. submitJob pays the fee.
Encrypted relay frames decrypted live with the session key.
The SDK is tested with real LCAI before each release. These ran the same code path you'd call.
decryptedDid you know there is a type of jellyfish called the 'Upside-Down Jellyfish' that actually swims on its back, using its tentacles to catch prey and defend itself from predators?
decryptedDid you know that the deepest part of the ocean, the Mariana Trench, is so deep that if you were to drop Mount Everest into it, its peak would still be more than 1 mile underwater?!