Superseed is where open-source contributions become on-chain rewards. Submit PRs, let AI judges score your work, and earn tokens for code that matters.
Three steps. No gatekeepers. Your code speaks for itself.
Push code to any participating repository. Superseed hooks into your workflow seamlessly through GitHub integration.
Multiple AI models independently evaluate complexity, impact, and quality. Consensus-driven scoring eliminates bias and gaming.
Scores are recorded on-chain. Tokens are distributed proportionally. The better the code, the greater the gravitational pull of your rewards.
Every design decision serves one goal: fair, transparent, ungameable code evaluation.
Multiple frontier LLMs score every PR independently. Cross-model consensus ensures no single model's bias dominates.
Every score, every rationale, every token distribution is recorded immutably. Audit anything, anytime.
Sybil detection, contribution fingerprinting, and anomaly scoring ensure rewards go to genuine builders, not manipulators.
A brilliant one-liner fix is worth more than a thousand lines of boilerplate. Context-aware evaluation rewards impact, not volume.
Every scored PR gets a structured AI review: what worked, what could improve, and how it compares to similar contributions.
Builders earn dynamic ELO ratings based on contribution quality over time. Rise through the ranks as your code proves itself.
Every great protocol starts with a single commit. Make yours count.