Nexez — Agentic Commerce

Agentic-commerce MCP: search across merchants, fetch structured offer manifests, and dry-run checkout + negotiation before buying.

GitHub

Documentation

Nexez

Pages AI agents can actually book from.

Nexez is the infrastructure layer for agentic commerce. Businesses create clean, structured agent-ready pages 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 pages 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

git clone https://github.com/yourusername/nexez.git
cd nexez
npm install
cp .env.example .env.local
# Fill in your Supabase, Stripe, and LLM keys
npm run dev
Visit http://localhost:3000 to start creating agent pages.

📋 How It Works

Connect your offers (Calendly, Stripe, Shopify, CSV, or manual)
Define rules per offer (pricing floors, availability, scope limits)
Publish – on Nexez subdomain or your own custom domain
Agents discover & negotiate – fixed offers book instantly, negotiable offers go through smart rules + LLM
Get paid – platform commission is automatically deducted, rest goes to you


💰 Revenue Model

Subscription: Free, Launch, Pro, Scale, Enterprise (contact sales for Enterprise)
Platform Commission: Applied to every successful agent-driven transaction (configurable per plan, including Free)
Business owners are the Merchant of Record for transactions


🛣 Roadmap

 Agent detection & analytics
 Negotiations inbox
 LLM pluggable architecture
 Full Smart Rules Engine + LLM auto-actions
 Persistent negotiation UI + structured JSON for agents
 Advanced escrow & dispute handling
 Marketplace directory of agent pages
 OpenClaw / autonomous agent workflow distribution


🤝 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](LICENSE).

Built with ❤️ for the agent economy