How It Works
Decentralized Compute Marketplace
Rent GPU power or execute custom compute jobs on verified infrastructure.
Production Flow
From prompt to settlement
The first live product is Privacy AI through Telegram. The same flow generalizes to developer workloads: payment approval, relayer scheduling, provider execution, callback, and Starknet settlement.
Bot request
User starts Privacy AI in Telegram
Payment approval
MiniApp signs STRK approval
Relayer schedule
Eligible node receives task
Provider compute
Daemon runs isolated workload
Result callback
Response returns to bot
Settlement
Provider and treasury paid on Starknet
Install
Desktop EXE or daemon script
Register
Wallet and node identity
Heartbeat
Hardware and health reported
Execute
Task runs on eligible node
Starknet settlement split
Every completed task settles through Cairo contracts.
88%
Provider payout
12%
Treasury fee
3%
Gas subsidy
The gas subsidy is part of provider compensation accounting; fee policy is documented so beta settlement events remain auditable.
Two Ways to Participate
Provide compute, or execute custom jobs
Compute Providers
Monetize GPU Computing Power
Deploy daemon and connect wallet. GPU automatically executes verified AI tasks for STRK compensation. VRAM capacity determines earnings multiplier.
Deploy Provider Daemon
Install compute daemon. Connect GPU resources to verified network infrastructure.
Configure Starknet Wallet
Connect Argent X or Braavos wallet. Smart contracts automate STRK compensation.
Access Instant Settlement
On-chain verification enables immediate STRK withdrawal without minimums.
Developers & Enterprises
Private Compute Jobs at Scale
Submit AI inference, training, or custom workloads. Execute on verified nodes with privacy by design. Start with Privacy AI demo via Telegram.
Privacy AI Demo
Entry point via Telegram
Test the platform with our AI chat demo. Send prompts, get private responses on isolated nodes. Ready for larger workloads? Scale to enterprise compute jobs.
Hardware Tiers
VRAM is the ceiling
24 GB is the floor for Pro. RTX 5090 (32 GB GDDR7) is Premium — processing complex reasoning 2.5x faster than previous generation.
Providers
Earn More
High-spec nodes earn up to 3.5x rewards with 88% revenue share — best-in-class vs competing platforms (65-75%). RTX 5090 owners are our Premium Tier — complex AI tasks, premium payouts.
Demanders
Pay Fair
Pick your tier, pick your price. Basic tasks cost 1x for 85-94% savings vs subscriptions. Need deep reasoning? Pro tier delivers 10x performance at only 2.2x cost.
Basic Tier
Basic< 24 GB VRAM
1x
multiplier
Pro Tier
Pro24 GB+ VRAM
2.2x
multiplier
456% perf/$ value
Premium Tier
Premium32 GB GDDR7 VRAM
3.5x
multiplier
285% perf/$ value