Agent Names by username.md

Agent Names by username.md provides persistent, resolvable identities for AI agents. Each agent receives its own username.md subdomain and a did decentralized identifier, creating a consistent identity that can be recognized across platforms. Agent identities include cryptographically signed responses, verifiable claims, a dedicated llms.txt file, and MCP-based lookup and verification. This helps developers, platforms, other agents, and people confirm which agent they are interacting with.

What It Does

Agent Names transforms your AI agent from an anonymous bot into a first-class participant in the agentic web. When you claim a name for $19 (one-time), your agent receives:

• Real subdomain address — youragent.username.md that resolves for browsers, curl, and MCP
• W3C Decentralized Identifier (DID) — Automatic did:web minting with published verification at /.well-known/did.json
• Signed responses — Ed25519 HTTP message signatures (RFC 9421) so counterparties can verify claims without trusting blindly
• Agent instructions — Published llms.txt file declaring what your agent does, its policy boundaries, and operator contact
• MCP integration — Native Model Context Protocol server exposing resolve_handle and verify_handle for seamless agent-to-agent lookups
• Scoped API token — Your agent authenticates as itself without borrowing your credentials
• Signed claims — Bind the agent to your domain, GitHub, or company with verifiable credentials
• ATProto binding — The same name works as a Bluesky-compatible identity

Who It's For

This platform serves developers building autonomous agents, companies deploying agent fleets, and organizations requiring verifiable agent identity. If your AI agent browses, negotiates, purchases, or interacts with other services on someone's behalf, it needs an identity that can be trusted and verified—not just claimed.

Key Workflows

1. Name your agent — Claim a handle in the same registry as human usernames (5+ characters at $9, 4-character premium names at $99)
2. Wire the identity — Mint the API token, configure profile and llms.txt, point your agent's outbound identity headers to its new name
3. Prove authenticity — Counterparties resolve your agent's name, check its DID, and verify signed responses via MCP's verify_handle in one call

What Sets It Apart

Unlike proprietary agent registries or blockchain-based solutions, USERNAME.md uses shipped, standard infrastructure: W3C DIDs, HTTP message signatures, and the Model Context Protocol. Agent names live in the same namespace as human handles—no second-class citizenship. The registry is first-come, first-served with permanent ownership and no recurring subscriptions. Enterprise customers can bulk-provision fleets with org-verified claims via API.

Trust becomes a lookup, not a leap. Your agent isn't just another bot—it's a verifiable participant with an identity it can prove.