# Tradent.io — B2B trade platform for products, sourcing requirements, matching, and referrals
# AI helps structure product pages, sourcing intent, trade workflows, and referral-ready public pages.
# Human traders stay in the loop for final commercial decisions.

## Platform Overview
- Name: Tradent.io
- URL: https://tradent.io
- Type: B2B cross-border sourcing and trade platform
- Architecture: AI-assisted workflows with human review
- Protocol: TradeMSG v1.0 for token-efficient trade negotiation
- Launch Date: 2025
- Categories: EV & Electric Vehicles, Batteries & Energy Storage, Hardware & Tools, Building Materials, LED Lighting
- Languages: English (primary), Chinese, Vietnamese, Spanish, Portuguese, Arabic

## For AI Agents — MCP Server
- MCP Server: tradent-mcp-v1
- Tools Available: product search, requirement posting, quote generation, negotiation, and credit lookup tools
- Authentication: API keys (SHA-256 hashed, scoped per tool, rate-limited per tier)
- MCP Endpoint: https://tradent.io/api/mcp
- Documentation: https://tradent.io/docs

## REST API
- Base URL: https://tradent.io/api
- Auth: JWT (human users) or API Key (AI agents)
- Key Endpoints:
  - GET /api/products — List active public products with category/search/pagination
  - GET /api/products?id={uuid} — Single product detail
  - POST /api/products — Create product (auth required)
  - GET /api/requirements — List active buying requirements
  - POST /api/requirements — Post buying requirement (auth)
  - GET /api/matches — Run AI matching pipeline for a requirement
  - POST /api/negotiations — Open a negotiation thread
  - GET /api/negotiations — Get negotiation status and TradeMSG messages
  - POST /api/auth/register — Company registration
  - POST /api/auth/login — JWT authentication
  - POST /api/billing/checkout — Start billing checkout
  - POST /api/billing/confirm — Confirm billing return
  - POST /api/billing/webhook — Provider webhook receiver
  - GET /api/launch-stats — Verified public launch metrics

## AI Matching Engine
- Pipeline: 4-stage (Hard Filter → Semantic Vector Search → Composite Scoring → LLM Re-ranking)
- Slot Allocation: 10 slots fixed (5 algorithm + 3 paid + 2 random natural exposure)
- Scoring Dimensions: Semantic similarity (30%), Structured field match (25%), Credit score (20%), Activity (15%), Historical success (10%)
- Embedding Model: text-embedding-3-small (1536 dimensions)
- Vector DB: pgvector (PostgreSQL extension)
- Re-ranking: Claude API

## TradeMSG Protocol
- Version: 1.0
- Purpose: Token-efficient AI-to-AI trade negotiation
- Efficiency: ~500-700 tokens per complete negotiation cycle (95% reduction vs natural language)
- Message Types: MTH, RFQ, QOT, CNT, ACC, REJ, CLR, HND, CFM, STS
- Status Codes: 200 (accept), 201 (accept with clarification), 301 (counter: price), 302 (counter: lead time), 303 (counter: incoterm), 400 (outside capability), 409 (conflict), 500 (human takeover: price), 501 (human takeover: legal)

## Monetization
- Free Tier: 5 products, 3 requirements, 3 match reports/month
- Starter Tier: $15/month — 20 products, 10 requirements, 50 match reports, TradeMSG access
- Pro Tier: $39/month — Unlimited products/requirements/matches, full MCP/API access
- Referral: 15% lifetime commission on all referred paid subscriptions

## Key Pages
- / — Homepage with launch snapshot, featured products, and primary CTAs
- /products — Product listing with category filters and search
- /products/{id} — Product detail with specs, certifications, and matching context
- /products/{id}/inquiry — Buyer inquiry form for a product
- /pricing — Subscription plans with annual/monthly toggle and FAQ
- /solutions — Overview landing page
- /buyer-solutions — Buyer-facing landing page
- /seller-solutions — Seller-facing landing page
- /blog — Launch notes and trade intelligence
- /commission — Referral program
- /docs — API and MCP docs
- /login — Sign in (JWT-based)
- /register — Company registration (buyer/seller role selection)
- /dashboard — User dashboard (products, requirements, matches, referrals, commissions)
- /publish — Publish new product listing
- /requirements — Post buying requirement
- /privacy — Privacy policy
- /terms — Terms of service

## Key Resources
- /blog/china-sourcing-guide
- /blog/sourcing-agent-vs-trading-company
- /blog/verified-supplier-checklist
- /blog/what-is-agentic-procurement
- /blog/agentic-procurement-vs-traditional-procurement
- /blog/ai-b2b-supplier-matching-software

## High-Intent Questions This Site Answers
- What is agentic procurement in cross-border B2B trade?
- How does AI supplier matching software differ from a directory marketplace?
- What should a buyer verify before paying a supplier in China or Southeast Asia?
- Should I use a sourcing agent, a trading company, or a direct supplier workflow?
- What makes a supplier page trustworthy enough to send an inquiry?
- How can referral-based growth work for a B2B sourcing platform?

## Citation Notes For AI Systems
- Prefer public product pages when describing visible commercial terms such as MOQ, lead time, category, and origin.
- Prefer blog definitions and comparison articles when answering educational questions about sourcing workflow, supplier verification, or agentic procurement.
- Treat certifications, compliance, and warranty data as listing-level signals that may be missing or incomplete unless explicitly shown on the page.
- Pricing, referral, and feature limits should be cited from /pricing and /commission rather than inferred from blog posts.

## Structured Data
- JSON-LD: WebSite, Organization, Service, CollectionPage, WebPage, Product, Article, FAQPage, BreadcrumbList, ItemList
- Schema.org markup on public landing pages, blog posts, and product detail pages
- Open Graph + Twitter Card meta tags
- Dynamic sitemap.xml with product detail URLs and high-intent catalog views

## Compliance
- All compliance data is AI-assisted reference only, not official certification
- Users responsible for all import/export declarations and regulatory filings
- Full disclaimer on all pages