No verifiable identity
Anyone can fake an AI agent on-chain. Zero cryptographic proof of authenticity, no standard to verify an agent's origin.
CODEX·ID is the decentralized identity & trust registry for autonomous agents — one NFT, three on-chain registries, full cryptographic provenance.
Billions of autonomous agent interactions will hit the chain in 2026–2027 — today they run on blind trust, with no provenance, no track record, and no off-switch.
Anyone can fake an AI agent on-chain. Zero cryptographic proof of authenticity, no standard to verify an agent's origin.
No way to query an agent's track record. Blind trust leads to massive financial risk — no on-chain feedback or scoring system.
Agents can't be paused, audited, or terminated on-chain. Zero accountability once they're deployed into the wild.
CODEX·ID gives every agent a permanent on-chain identity, then layers trustless reputation and third-party validation on top — a full trust stack, queryable by anyone.
getSummary() queryable by anyone, anytime.validationResponse() callable on demand.One NFT / Three registries / Full trust stack.
A single call to register() mints the identity NFT — where agentId = tokenId. Transfer the NFT and you transfer the identity, reputation, and every byte of metadata with it.
Attestations, BAP-578 lifecycle state, and the permanent agent wallet address — all bound to one token.
The client feedback loop. Scores written on-chain, appendResponse() for the agent, getSummary() for everyone.
Independent third-party scoring with getSummary() and validationResponse() callable on-chain.
The NFT is the agent. Transfer the token, transfer the whole identity.
The first project to implement ERC-8004 + BAP-578 together on opBNB. No custom forks, no lock-in — every layer is an open, audited standard.
Trustless Agents standard — the three-registry pattern that CODEX·ID is built on.
28 / 28 requirementsNon-Fungible Agent — lifecycle & metadata standard.
17 / 17 coreNFT backbone — each agent is a unique on-chain token.
Full supportTyped structured-data signing for agent wallet verification.
Full supportSmart-contract wallet signature validation support.
Full supportAll OpenZeppelin standards deploy without modification — zero custom tooling required.
Not a mock-up. The full stack is deployed and verified on opBNB Testnet, with real transactions you can inspect on the explorer right now.
Storing every feedback, attestation, and validation on-chain only works if it's cheap, EVM-native, and built where the ecosystem is heading. opBNB is all three.
Every feedback, attestation, and validation stored on-chain affordably — at massive agent scale.
ERC-721, EIP-712 and ERC-1271 deploy without modification. All OpenZeppelin standards just work — zero custom tooling.
BAP-578 is a BNB Application Proposal. We build exactly where the ecosystem is heading — alignment, not adaptation.
Any protocol that grants an AI agent power — over funds, deals, or listings — needs to know the agent is who it claims to be, and has the record to back it.
Verify an AI trading agent's identity and track record before granting permissions to manage funds.
getSummary()verifyAttestation()Buyers check an AI sales agent's reputation and attestations before engaging — scores queryable by any wallet, no intermediary.
giveFeedback()getSummary()List only verified agents using CODEX·ID as the trust gatekeeping standard — onboard with register(), screen with validationRequest().
register()validationRequest()CODEX·ID is built by a team already running battle-tested agentic AI in production, with tier-1 cloud partnerships and global infrastructure.
Orvium.ai — a live, production agentic-AI platform already in the wild.
Strategic partnerships with AWS and Google Cloud.
Physical offices in Dubai and Santiago supporting international operations.
A dedicated Latin-American team driving localized innovation and growth.
Recent presence at Function1 and the Doers Summit Dubai.
“In a world where AI agents transact autonomously, CODEX·ID is the trust layer that makes it safe.”