// Front matter
ParalleliX Documentation
The marketing site sells the idea.
// One document
The marketing site sells the idea. This document is the specification. It explains how ParalleliX works in the order it works, with every protocol primitive named, every economic parameter locked, and nothing rounded off for effect. Builders, operators, ParalleliX AI users, and holders read from the same source. There is no marketing version, no investor version, no whitepaper-grade approximation. One document. You are reading it.
// Document parameters
Chapters
15
I through XVI · Arabic-numbered
Subsections
64
Deep-linkable · per-section anchors
Reading time
2h
118 minutes end-to-end
Revision
v0.3
Updated 2026-05 · live
// Starting points
Four entry points. One document.
Each entry drops you on the chapter that builds the fastest mental model for what you're doing. No wrong door.
- // 01First-time readers
Read the thesis
The shape of the network, before the mechanics. Why parallel compute needs its own coordination layer.
- // 02Engineers
Trace a request
How a ParalleliX AI inference request moves through the network: routed whole to one operator node, executed, proved, validated, settled. Parallelizable workloads add segment and aggregate on the planned marketplace.
- // 03Operators · partners
Run a node
Bare metal to first paid request: install, register, accept dispatches, earn $PRLX for uptime.
- // 04Holders · analysts
Inspect the token
Supply, distribution, tax, permissionless staking, ownership scope. The full mechanics of $PRLX.
// Reference shelf
Quick-lookup pages
Three reference surfaces the rest of the document leans on. Keep one open in a separate tab while you read.
// Begin
Start with First Principles.
The shape of the network, before the mechanics. Five sections, ten minutes. Every other chapter assumes you've read it.