Loading...
Loading...
Discovery and trust bootstrapping for the Internet of AI agents
A new indirection layer that incorporates forty years of hard-won lessons about global DNS infrastructure, governance, trust bootstrapping, and financial sustainability—applied using distributed and decentralized systems.
AgenticDNS is inspired by the NANDA Architecture developed at MIT Media Lab. Each layer serves a distinct purpose, enabling rapid evolution of capabilities and trust signals without touching the global index.
A lightweight, globally replicated index maps agent identifiers to metadata references, not endpoints. Each AgentAddr record is ≤120 bytes, signed, and cacheable.
Capabilities, endpoints, and trust assertions live in signed metadata documents (JSON-LD, W3C Verifiable Credentials) and evolve independently of the index.
Resolution decisions occur locally, under enterprise policy, at machine speed. Supports static endpoints, rotating pools, and adaptive routing.

Agents are ephemeral, mobile, and operate across organizational boundaries at machine speed. DNS was designed for static hosts and human timescales.
Current certificate revocation (CRL/OCSP) cannot keep pace with millisecond-level agent revocations. AgenticDNS enables sub-second revocation via VC-Status-List mechanisms.
TLS certificates prove domain ownership, not agent behavior or capability. AgentFacts provides machine-verifiable trust and capability signaling.
Agents need verifiable claims about capabilities, security posture, and compliance—not just proof of domain control. AgenticDNS addresses trust+capability signaling for autonomous agents.
DNS continues to resolve hosts and domains. AgenticDNS introduces a parallel indirection layer that resolves agent authority, capabilities, and trust—concerns DNS was never designed to handle. This separation enables:
AgenticDNS evolves from intranet to extranet to internet, just as networking itself grew from local networks to global connectivity.
Single-organization deployment with tight governance and highest control. Enables internal agent discovery and capability negotiation.
Multi-organization federation with negotiated trust exchange. Enables secure agent collaboration across partner boundaries.
Global public discovery and trust bootstrapping. Enables open agent ecosystems with verifiable identity and capability assertions.
Have questions or want to learn more about AgenticDNS?
Email: info@agenticDNS.ai