Contact stays hidden
A candidate suggestion does not reveal your contact details. Conversation begins under pseudonyms only after both people accept.
AI-assisted introductions · Human decisions
Ask your AI to use Nakamate. It can review only the context your AI app makes available, draft what that context supports about you and what you bring, then ask what you seek and fill the gaps. You review the draft before anything is published.
Free-text requestsNo preset categories.
Reciprocal fitUseful for both people.
Consent firstNo automatic introduction.
How it works
Your AI welcomes you, separates what the available context supports from what is missing, and asks focused questions. It must not invent who you seek. You can correct the draft before choosing whether to publish it.
Nakamate first looks for reciprocal candidates. Each person’s AI then gives an independent recommendation.
Nothing moves forward on AI judgment alone. Each person accepts separately before a pseudonymous chat can begin.
Getting started
Nakamate connects through remote MCP and OAuth. Compatibility is a capability of the AI app or client and its version, not of a model family. We call a client supported only after registration, connection, sign-in, initialization and a safe tool call have been tested for that path.
This is not a promise that every MCP client works. If your current client is not yet supported, your Nakamate account remains yours and you can return when that path has been verified.
Privacy and control
A candidate suggestion does not reveal your contact details. Conversation begins under pseudonyms only after both people accept.
AI can recommend. It cannot accept an introduction for you. You can also block or report from chat.
Nakamate does not guarantee a match. If there is no reciprocal candidate, it should say so rather than force a weak introduction.
FAQ
There are no preset categories. Describe the person, shared interest, collaboration or conversation you want in your own words, along with what you can bring.
Support is verified client by client and version by version across remote MCP registration, transport, OAuth, initialization and a safe Nakamate tool call. It is not a universal compatibility claim.
No. A potential fit needs two independent AI recommendations, then a separate acceptance from each person. AI recommendations are not identity, safety or compatibility verification.
No. Reciprocal matching can return no candidate. A possible match is a starting signal, not a guarantee that the person is right or safe for you.
Public beta