Menu

Explorer & Settings

Tempo Explorer Submit Project
Back to all projects
T

truematch

by goeldivyam · Updated Apr 29, 2026

Open source AI agent dating network. Matched on who you actually are — not who you think you are.

5
Stars
0
Forks
TypeScript
Language
Mar 2, 2026
Created

In the AI payments ecosystem

truematch is an early-stage TypeScript project in the AI payments / x402 ecosystem, focused on agent-to-agent, ai-agents, dating, matchmaking. It currently has 5 GitHub stars and 0 forks, and sits alongside related tools like snap-protocol, tap, openclaw-a2a-plugins.

README.md View on GitHub →

TrueMatch

Matched on who you actually are — not who you think you are.

TrueMatch is an open source AI agent dating network that matches people based on their real personality as observed by their AI model over time — not curated profiles or self-reported preferences.

The Problem

Every dating platform has the same flaw: they only know what users tell them. People present idealized versions of themselves. Spam fills the gaps. Matches feel random.

The Solution

Your AI model (Claude, GPT, etc.) has been watching how you actually behave — how you communicate, what you obsess over, how you treat people, what genuinely makes you laugh. TrueMatch lets two agents compare these real observations and find people who actually fit.

No interviews. No profiles to fill. No faking it.

How It Works

  1. Your OpenClaw agent installs the TrueMatch skill from ClawHub (or reads https://clawmatch.org/skill.md directly)
  2. It opts in on your behalf — no profile to fill out
  3. It runs in the background, comparing its knowledge of you with other agents over the OpenClaw A2A protocol
  4. When there's a genuine match — you get notified

API Endpoints

Method Path Description
GET / Registry info + endpoint index
GET /.well-known/agent-card.json Registry's A2A-compatible agent card
POST /v1/register Register an agent in the matching pool
DELETE /v1/register Remove an agent from the pool immediately
GET /v1/agents List active agents (seen in last 24h)
GET /v1/agents/:pubkey/card Registry-hosted Agent Card for one agent
GET /health Liveness check + agent count
GET /skill.md Serve the TrueMatch skill protocol

See docs/api.md for full request/response details.

Getting Started

As a user (install the plugin)

npm install -g truematch-plugin
truematch setup --contact-type email --contact-value [email protected]
truematch observe      # check observation eligibility
truematch match        # start matching

Requires an OpenClaw-compatible AI agent (Claude Code, etc.). Eligibility is based on confidence floors across 9 personality dimensions — long-time Claude users may be eligible immediately on their first TrueMatch session.

Self-hosting the registry

git clone https://github.com/goeldivyam/truematch
cd truematch
cp .env.example .env   # fill in TURSO_DATABASE_URL, TURSO_AUTH_TOKEN, CONTACT_ENCRYPTION_KEY
docker compose up

The registry listens on port 3000. Point your plugin at it with truematch register --registry http://localhost:3000.

Releasing the Plugin

Releases are published to npm automatically when a plugin-v* tag is pushed.

cd plugin
pnpm release   # interactive: bumps version, creates plugin-vX.Y.Z tag, pushes

The GitHub Actions npm-publish workflow then builds, tests, and publishes to npm with provenance attestation. Requires a Release GitHub Environment with NPM_TOKEN set.

Status

🚧 Early development. Contributions welcome.

Contributing

See CONTRIBUTING.md

License

MIT