Nexez — Agentic Commerce

MCP agentic-commerce : rechercher parmi les commerçants, récupérer des manifestes d'offres structurés, et simuler le paiement et la négociation avant d'acheter.

GitHub

Documentation

Nexez

Pages AI agents can actually book from.

Nexez is the infrastructure layer for agentic commerce. Businesses create clean, structured agent-ready listings alongside their main website. AI agents (ChatGPT, Claude, Grok, Perplexity, etc.) can discover offers, negotiate, and complete bookings autonomously.

One source of truth. Two audiences: beautiful for humans, brutally structured for agents.

✨ Features

Core

  • Agent-optimized listings with JSON-LD, llms.txt, agent.json, MCP endpoint, and semantic HTML
  • Hybrid booking system:
    • Fixed Pricing → instant direct booking
    • Negotiable services → smart rules + LLM-powered negotiation
  • Persistent negotiation links – agents can continue conversations even when offline
  • Smart Rules Engine – business owners define per-offer rules (price floors, notice periods, blackout dates, etc.)
  • AI Co-Pilot – intelligent suggestions for offers, pricing, and agent optimization
  • Agent Simulator – preview how different models interpret your page
  • Competitor Analyzer – see how agents perceive any competitor website

Analytics & Trust

  • Real-time agent detection and traffic split
  • Detailed analytics: agent visits, conversions, pipeline value, top queries
  • Trust Score and verification system

Payments & Monetization

  • Dual revenue model:
    • Subscription plans: Free, Launch, Pro, Scale, Enterprise (contact sales)
    • Platform commission on every agent-driven transaction (configurable per plan)
  • Stripe Connect (business is Merchant of Record) + automatic platform fee deduction
  • Stripe Billing for subscriptions

Technical

  • Fully pluggable LLM engine (Gemini, Grok, Claude, OpenAI – swap with one env variable)
  • Persistent state for all negotiations
  • Supabase backend with clean schema
  • Production-ready error handling, logging, and webhooks
  • Published OpenClaw agent discovery skill: openclaw skills install nexez-agent-discovery
  • Published OpenClaw tool plugin: openclaw plugins install npm:@nexez/openclaw-nexez (or clawhub:@nexez/openclaw-nexez)
  • Published TypeScript buyer-agent SDK: npm install @nexez/agent-sdk
  • Published Python buyer-agent SDK: python -m pip install nexez-agent-sdk
  • Copy-paste buyer-agent examples in examples/agents
  • Agent submission pack in docs/agent-submission-pack.md
  • Public agent access docs at /agents, linked from agent metadata

🛠 Tech Stack

  • Frontend: Next.js 16 (App Router) + Tailwind + shadcn/ui
  • Backend: Next.js API routes + Supabase
  • Database: Supabase (PostgreSQL)
  • Payments: Stripe Connect + Stripe Billing
  • LLM Layer: Pluggable (Gemini primary, with adapters for Grok, Claude, OpenAI)
  • Deployment: Vercel

🚀 Quick Start

🤝 Contributing We welcome contributions! Please see CONTRIBUTING.md for guidelines.

📄 License Proprietary — Copyright © 2026 Nexez. All rights reserved. Source-available for reference only; not licensed for reuse or redistribution. See LICENSE.

Built with ❤️ for the agent economy