Documentation

Everything you need to buy, trade, and deploy tokenized GPU compute. From first token to production cluster.

Quick Start

Go from zero to GPU session in 5 minutes.

API Reference

Full endpoint documentation with examples.

SDKs & Libraries

Python, Node.js, and Go client libraries.

Getting Started

Platform Overview

Understand the tokenized compute model and how Lubbock Cloud works.

Create an Account

Set up your wallet connection and fund your first token balance.

Your First GPU Session

Redeem tokens and launch an interactive GPU session in under 5 minutes.

Compute Tokens

Token Mechanics

How tokens map to GPU-hours, memory tiers, and priority levels.

Buying & Selling

Trade tokens on the marketplace with market or limit orders.

Redemption

Convert tokens into live sessions, API credits, or batch jobs.

Vaults & Staking

Lock tokens to earn yield, discounts, and priority access.

API Reference

Authentication

API key management, scopes, and rate limits.

Compute API

Launch jobs, check status, and manage GPU sessions programmatically.

Token API

Query balances, submit orders, and manage vault positions.

Webhooks

Real-time event notifications for job completion, order fills, and vault maturity.

ROCm Services

LLM Inference

Deploy open-weight models with vLLM on MI300X hardware.

Embeddings API

High-throughput text and image embeddings at scale.

Fine-Tuning

LoRA, QLoRA, and full-parameter training pipelines.

Vector Database

GPU-accelerated FAISS vector search at billion-scale.

Enterprise

Cluster Configuration

Dedicated GPU clusters with InfiniBand networking.

SLA & Support

Enterprise SLA options, 24/7 support, and dedicated account management.

Compliance

SOC 2, data residency, and security documentation.

Example: Launch an Inference Job

import { LubbockClient } from '@lubbock/sdk';

const client = new LubbockClient({
  apiKey: process.env.LUBBOCK_API_KEY,
});

// Redeem 4 LUB-MI300X tokens for an inference session
const job = await client.compute.create({
  gpu: 'MI300X',
  durationHours: 4,
  jobType: 'inference',
  config: {
    model: 'meta-llama/Llama-3-70B',
    quantization: 'fp16',
    maxBatchSize: 32,
  },
});

console.log(job.id);       // "job-abc123"
console.log(job.endpoint); // "https://compute.lubbock.cloud/job-abc123"
console.log(job.status);   // "provisioning"