The trust layer infrastructure for AI agents onchain identity, registry, and economic primitives for the open agentic web built on Arbitrum .
Live link : https://agent-multiverse.vercel.app/
IN this folder i have shared the Demo video & pitch video for the project find the accessible link here :
DEMO : https://drive.google.com/file/d/1LwycfGOHqpuE5gdqcub8qzKlORvGRTal/view?usp=sharing
PITCH : https://drive.google.com/file/d/1LwycfGOHqpuE5gdqcub8qzKlORvGRTal/view?usp=drive_link
Agent Multiverse is a permissionless trust infrastructure for AI agents and MCP services, built natively on Arbitrum. It solves a specific problem: as AI agents proliferate, there is no shared, verifiable layer for discovering, trusting, and coordinating between them. We provide that layer onchain.
Three composable onchain primitives that any protocol, wallet, or AI platform can plug into:
| Primitive | What It Solves |
|-----------------|----------------|
| Agent Passport | Permanent, portable identity for every agent (ERC-721) |
| Agent Registry | Verifiable on-chain directory of AI services with audit trail |
| Audit Bounties | Economic incentive layer that rewards independent verification |
How It Works
Three steps from anonymous agent to trusted service:
01 REGISTER
Developer publishes their AI service on-chain
→ name, endpoint, repo URL, description
02 AUDIT
Independent auditors review and submit reproducible reports
→ service gains isVerified = true on-chain
03 EARN
Sponsors post token bounties to incentivize audits
→ auditors claim bounties when verification passes
No central authority. No manual review. Permissionless by design.
Hackathon Period — Now (Complete)**
Week 1- Three core contracts deployed and verified
Week 2- Frontend & UI Polishing
Week3- Next.js dashboard + CLI
Final days- Arbitrum Sepolia live & testing of the infra
Looking to start raising funds to move on to mainnet phase