I am building an open-source DePIN protocol built to able high-end hardware to start node in a private and transparent compute layer.
I want to replace “black box” centralized cloud models with a verified, peer-to-peer network where infrastructure is treated as a public good.
Currently implemented:
lightweight daemon to seamlessly register node to network, detect your hardware and get matched jobs.
Pay-per-Task - fully work assumption system. enables a true micropayment economy where providers are paid in STRK immediately upon the successful validation of a compute task and providers get paid with STRK, not other token.
Privacy: local execution environments (WSL2/Docker) to ensure data remains isolated. Starknet’s ZK-rollups provide the cryptographic proof of work without exposing the underlying data.
Telegram first Privacy-AI - @smainer_ai_bot that you can pay per prompt with STRK
Fully working relayer: Event-driven architecture.
Pre-Launch Feedback Requested:
I’m finalizing the protocol before the official public launch. I am looking for technical feedback:
Security Review: Feedback on the Cairo escrow contract logic and Relayer authentication.
Repo Review: I have a background in enterprise software-engineering but I used AI tools for much of the build and I’m not a security expert.
General feedback: Just hear the community general opinion.
I am looking for a peer review while the repo is still private before the public launch.
We are building open-source infrastructure for private AI compute on Starknet: a network where users can submit Privacy AI jobs, providers can run compute nodes, and settlement happens in STRK with transparent on-chain coordination.
Our current focus is simple:
Make the stack inspectable.
Make provider onboarding practical.
Make private compute cheaper and easier to access.
Make every core system understandable enough for builders to verify, improve, and extend.
What is live or in progress:
Public frontend for compute jobs, provider onboarding, and network status
Desktop app for Windows provider nodes
Relayer infrastructure for job scheduling and provider coordination
Starknet contracts for registry, escrow, and proof flows
Telegram access for lightweight Privacy AI workflows
Open documentation for architecture, setup, and operations
What we are planning next:
Cleaner public repositories and contribution paths
Better provider diagnostics and node reliability tooling
More transparent live network metrics
Stronger verification flows between relayer, provider, and Starknet
Expanded docs for builders who want to run, audit, or contribute
More practical examples of Privacy AI jobs running through the network
Smainer is infrastructure that should be visible, testable, and useful to the people building with it.
If you are working on AI infrastructure, Starknet, decentralized compute, provider networks, or privacy-preserving workloads, we are opening the door.