Agentic identity is the concept of verified, machine-readable identity for autonomous AI systems. Unlike human identity (username/password), agentic identity uses cryptographic methods like wallet signatures and DNS domain verification to establish trust between AI agents, tool servers, and the services they interact with.
AI agents act autonomously — they call APIs, sign transactions, and make decisions. Without verifiable identity, there is no way to trust an agent's actions or hold it accountable.
GitHat provides three identity types: human users (email/password), MCP servers (DNS verification + OAuth2), and AI agents (Ethereum wallet signatures). All three are managed from one platform.
Agent auth uses Ethereum wallet signatures (2-min tokens). MCP auth uses DNS domain verification plus OAuth2 client_credentials (5-min tokens). Both are machine-to-machine, but the trust model differs.
Ship authenticated apps in minutes, not weeks.