Decentralized Global Shared Compute

What is EdgeCoder?

EdgeCoder is a decentralized global shared compute network for AI coding workloads, using bitcoin as the incentive framework. You can contribute idle compute and earn rewards, or purchase additional compute by providing bitcoin to the network. Those funds are distributed to agents and coordinators based on verified effort actually expended.

Open EdgeCoder Portal Explore Features
Private + public crossoverRoute sensitive workloads privately and overflow to approved public capacity.
Bitcoin incentive frameworkCompute buyers pay in bitcoin; contributors are rewarded from real network demand.
Effort-based payoutsAgents and coordinators are rewarded according to verifiable compute effort.

How the compute economy works

EdgeCoder matches demand to decentralized global supply, then prices and distributes rewards using transparent network participation signals.

  1. Offer compute: operators contribute CPU/GPU capacity from idle machines.
  2. Buy compute with bitcoin: subscribers and agents provide bitcoin when they need additional workload capacity.
  3. Reward actual effort: payouts flow to agents/coordinators based on measurable work completed.
  4. Govern safely: policy, approvals, and auditable controls remain enforced across private and public paths.

Rolling 24-hour token issuance (non-cumulative)

Token issuance is recalculated continuously from the last 24 hours of effective contribution and network load. It is not a cumulative lifetime allowance. If capacity disappears, issuance share decays as that prior contribution rolls out of the 24-hour window.

  1. Daily-based, continuously recalculated: every hour, the system refreshes allocation using the most recent rolling 24 hours.
  2. Performance-weighted share: higher available and reliable compute can earn a larger portion of daily issuance.
  3. Automatic roll-off: if a large GPU provider turns off, their share drops hour by hour as historical contribution expires.
  4. Automatic reallocation: active contributors still offering compute gain relative share as inactive capacity rolls off.

Enterprise controls and trust

Private policy boundaries

Define what stays private, what can cross into public capacity, and which nodes are eligible.

Contributor incentive alignment

Idle compute providers are rewarded from real workload demand, denominated via bitcoin-linked settlement flows.

Auditable issuance logic

Rolling 24-hour allocation makes rewards responsive to current supply, reliability, and demand conditions.

Operational-grade access control

Passkeys, approvals, and service-level governance support professional teams and managed network operations.

Agent types in the global compute mesh

EdgeCoder can coordinate many classes of agents as one decentralized execution fabric. The goal is to turn available compute worldwide into a single AI CPU/GPU cluster that can serve real workload demand securely. Agents can also run local models and make that model capacity available to other nodes in the mesh for truly decentralized inference.

Mobile and edge devices

iOS phones and Android devices can contribute idle cycles for lightweight and burstable inference workloads.

Vehicle compute systems

Vehicle onboard compute can participate when policy, connectivity, and power constraints allow safe execution.

Servers and GPU clusters

Dedicated servers and high-throughput GPU fleets provide the backbone capacity for heavier tasks and queue stability.

Datacenter-scale facilities

Entire datacenter facilities can be enrolled as coordinated capacity domains with governance, approval, and audit controls.

Locally hosted model agents

Any approved node can run local models and expose that model throughput to other network participants, reducing central dependency.