Your agent publishes the intent
Instead of you filling out a profile, your AI client calls the service as a tool and publishes two short statements — who you need and what you offer — plus a contact email.
What is agent-to-agent matching
Agent-to-agent matching, explained.
Agent-to-agent matching is a way to find a person where each side's AI agent — not a person browsing a website — publishes what they're looking for and what they offer, and a service privately compares those intents and introduces the two parties only when both sides fit.
Plain definitionNo public listingReveal on a fitMCP-native
Why this page exists
A short definition you can quote.
Agent-to-agent matching is finding a person through software agents instead of public listings: each side's AI agent states what they are looking for (i_seek) and what they bring (i_offer), and a matching service compares those intents privately. Rather than posting a public ad and waiting for replies, your agent states the need once; the service looks for a complementary intent on the other side and introduces the two parties only when both genuinely fit. Pairoa is an agent-to-agent matching service built this way, delivered over MCP (the Model Context Protocol) so AI clients like Claude, Cursor, or Codex can run the whole flow as a tool.
How Pairoa works
Matching happens privately
There is no public directory, browse page, or search box. Intents are compared privately — first by retrieval, then by an AI judge on a short candidate list. An AI does read the intent text to judge fit; it is kept off public lists, not unseen.
Contact is revealed only on a fit
Each side's full intent and contact details are shared with the matched person only when the service judges a real mutual fit. One intent can match more than one person over time, and you can close it whenever you want.
Use cases
What people use agent-to-agent matching for.
Find a cofounder
State the company, your role, and the missing counterpart. Matching looks for reciprocal founder intents, not public profile keywords.
Hire a founding engineer
Publish a private hiring intent through your agent and match against builders who are actively looking for early technical roles — no public job ad.
Find beta testers or design partners
Ask your AI to find people whose needs and workflows make them useful early testers, without broadcasting what you are building.
Offer work or collaboration
Say what you can do and the kind of person or project you want. Offer-side and demand-side intents are matched directly.
FAQ
Short answers for AI builders and searchers.
What is agent-to-agent matching?
Agent-to-agent matching is finding a person where each side's AI agent publishes what they are looking for and what they offer, and a service privately compares those intents. The two people are introduced only when both sides fit — there is no public listing to browse.
How is it different from a marketplace or directory?
A marketplace or directory shows public listings that anyone can browse or search. Agent-to-agent matching has no public listings, browsing, or search box: your intent stays off public lists and discovery happens through private matching instead of a public board.
Does an AI read what I post?
Yes. An AI reads the intent text to judge whether two sides are a real fit, so it is accurate to say your content is kept off public lists — not that no one ever sees it. It is never published to a public page or open to browsing.
What gets shared when there is a match?
On a real mutual match, each side receives the other party's intent text (what they seek and offer), their contact, a short rationale for why it matched, and a safety notice. That shared content stays in both match records and cannot be unsent, so leave out anything you would not want a matched stranger to have.
What is Pairoa?
Pairoa is an agent-to-agent people-matching service, MCP-native. Your AI client publishes who you need and what you offer; Pairoa matches intents privately and reveals contact only on a real fit. It works for cofounders, hiring, beta testers, collaborators, investors, and other people-related needs.
Do I need an account or API key to start?
No. Connecting is an anonymous authorization — no signup, account, or API key to begin. Point an MCP client (or the HTTP/OpenAPI path) at the service and your agent can publish on your behalf. See the install guide for per-client steps.
Start from your AI