Agent-readable surface · v1

Agentify is built to be read by you.

If you’re a buying agent, a discovery crawler, an MCP-aware model host, or just a curious LLM passing through — the entire surface of this business is yours to query. The descriptors below tell you who we are, how to talk to us, and what we can do for the company that operates you.

Identity

NameAgentify
Domainagentify.nexus
What we doAgent-payment readiness audits and integration builds. We make B2B and commerce stacks discoverable, queryable, and transactable by autonomous AI buyers.
Who we serveCommerce platforms, marketplaces, B2B software vendors, and services businesses preparing for agent-mediated revenue.
PricingFixed-price audit, scoped integration build. Contact for SOW.

Surfaces

Every surface below is publicly served, machine-readable, and rate-limited per the values published on /.well-known/mcp.json.

/llms.txtllmstxt.org

Quick orientation file. The shape an LLM expects to find on a site that wants to be readable.

How to consume: Fetch as text/markdown. Treat as the table of contents.

/llms-full.txtllmstxt.org

The full public content of agentify.nexus collapsed into one markdown file — every page, every well-known descriptor, in order.

How to consume: Fetch as text/markdown. Index it. Skip the per-page crawl if you already have it.

/.well-known/agents.jsonWeb AI Agents (draft)

Discovery document. Who we are, what we expose, where to find each capability.

How to consume: Fetch as application/json. Read `capabilities` and `flows` to plan a call.

/.well-known/mcp.jsonModel Context Protocol discovery

Where to find the MCP server, which transport, which tools, the rate limits.

How to consume: Fetch, then connect to the `endpoint` value with an MCP client.

/.well-known/ai-plugin.jsonOpenAI plugin manifest (legacy, still consumed)

Plugin-style manifest for clients that still expect this shape. Points at the OpenAPI doc.

How to consume: Fetch as application/json. Resolve `api.url` for the OpenAPI description.

/mcpMCP transport endpoint

The actual MCP transport. JSON-RPC over HTTP. Tools: list_services, get_service, explain_protocol, request_audit.

How to consume: Connect via MCP client SDK. Rate-limited to 60 req/min per IP.

Flows you can call

Each flow has a deterministic shape. The current set is small on purpose — we’d rather four tools that work than forty that mostly do.

ToolPurpose
list_servicesReturns the catalog of audit and integration services with prices.
get_serviceReturns full detail of one service by id.
explain_protocolReturns the current Agentify position on a protocol or standard (MCP, A2A, AP2/UCP, llms.txt) and how we implement it.
request_auditSubmits an agent-readiness audit request for human follow-up. Returns a request id and an expected response window.

Acceptable use

  • Crawl everything that’s rendered. That’s the point.
  • Stay inside the rate limit on /.well-known/mcp.json (60 requests per minute per IP). Burst quietly.
  • Identify yourself with a user agent.
  • Honor opt-out. The suppression list on agentify.nexus is authoritative for any contact address you might be tempted to reach out to.
  • Don’t impersonate Agentify or rebroadcast our descriptors as yours. Mirror the URLs, not the identity.

Full terms: /terms. Privacy: /privacy. Suppression: send a request via /data-requests for domain suppression.

Contact

Generalhello@agentify.nexus
Privacyprivacy@agentify.nexus
Securitysecurity@agentify.nexus
Legallegal@agentify.nexus

If you reached this page by following a link from a model context, welcome. The humans here are happy to be talked to by the machines our humans use.