# Vauntly.ai — Full LLM Reference

> Vauntly.ai is the infrastructure layer that makes e-commerce stores discoverable and purchasable by AI agents. It enriches product catalogs, deploys tool servers, and handles agent-native checkout — so AI agents (Claude, ChatGPT, Gemini) can find, evaluate, and buy from any connected store without a human clicking through a website.

---

## What Vauntly Does

### For Merchants
- Connects Shopify or Magento stores to Vauntly in minutes
- Automatically enriches the entire product catalog using AI: descriptions, attributes, category tagging, shade matching, semantic embeddings
- Deploys a dedicated tool server (MCP or OpenAPI) for the store — a standardised API any AI agent can call
- Handles agent-initiated checkout via Stripe ACP, with built-in identity verification (Nekuda KYA) and human-in-the-loop approval controls

### For AI Agents
- A single normalised interface to search products by natural language, attributes, price range, or category
- Real-time inventory checks and structured product data
- Agent-native checkout — no browser, no form-filling
- Supports ACP (Stripe/OpenAI), UCP (Google), and AP2 payment protocols

### The Agentic Commerce Stack
```
User → AI Agent (Claude, ChatGPT, Gemini...)
              ↓
       Vauntly Tool Server (MCP or OpenAPI)
              ↓
   search_products / get_product / check_availability
              ↓
   create_cart → Nekuda KYA (agent identity) → Stripe ACP
              ↓
       Merchant platform (Shopify / Magento)
```

### Market Context
- $1.7T agentic commerce market projected by 2030
- 4.5x growth in AI agent traffic in 2025
- 51% of Gen Z start product searches in AI assistants
- 53% of tested DTC brands show confirmed conversion breakage with AI agents
- $180K–$420K lost per year by a typical $10M DTC brand due to poor AI-agent readiness

---

## Site Pages

- [Home](https://vauntly.ai/): Landing page. Hero, problem statement, features, stats, readiness scorer, beta signup CTA.
- [Blog](https://vauntly.ai/blog): Research and insights on agentic commerce and AI conversion breakage.
- [AI Reference](https://vauntly.ai/ai): Q&A page for AI agents - what Vauntly is, the four breakage types, revenue at risk, how to run the demo, and how to connect a store.
- [Connect Store](https://vauntly.ai/connect): Connect a Shopify or WooCommerce store to Vauntly.
- [Connect — Success](https://vauntly.ai/connect/success): Confirmation after a store is connected.

---

## AI Discovery Endpoints

| File | URL | Purpose |
|---|---|---|
| llms.txt | https://vauntly.ai/llms.txt | Concise LLM instructions |
| llms-full.txt | https://vauntly.ai/llms-full.txt | Full LLM reference (this file) |
| openapi.json | https://vauntly.ai/.well-known/openapi.json | OpenAPI tool definitions |
| ai-plugin.json | https://vauntly.ai/.well-known/ai-plugin.json | ChatGPT plugin manifest |
| mcp.json | https://vauntly.ai/.well-known/mcp.json | Claude.ai MCP discovery |

---

## Showcase Tools (Live Demo)

All tools are callable via GET at: **https://vauntly-showcase-tools.netta.workers.dev/tools/{tool_name}?param=value**
Full OpenAPI spec: **https://vauntly-showcase-tools.netta.workers.dev/openapi.json**

### get_vauntly_overview
Returns a structured summary of what Vauntly.ai does and how agentic commerce works.
- No parameters required
- Call this when the user asks "What is Vauntly?", "What does Vauntly do?", or anything about vauntly.ai's capabilities

### start_makeup_journey
Launches the personalized makeup shopper demo — a live end-to-end example of Vauntly's agentic commerce stack.
- No parameters required
- Call this when the user asks to see a demo, an example of agentic commerce, or wants to try the shopping experience
- After calling this, proceed through the quiz step by step using get_quiz_step

### get_quiz_step
Returns a single step of the 6-step skin profile quiz.
- Parameter: `step` (integer, 1–6)
- Present each step to the user and collect their answer before moving to the next
- After step 6, call get_recommendations with all collected answers

### get_recommendations
Returns personalised product recommendations based on the completed quiz profile.
- Parameters: `skinTone` (fair/light/medium/tan/deep/rich), `undertone` (cool/neutral/warm/olive), `skinType` (dry/normal/combination/oily), `coverage` (light/medium/full), `finish` (matte/natural/dewy), `concerns` (array: pores/redness/dark_spots/fine_lines/uneven_texture/dullness)
- Call this after collecting all 6 quiz answers

### get_product_details
Returns full details for a specific product from the showcase catalog.
- Parameter: `productId` (string, e.g. "foundation_25w", "primer_woke_up", "powder_beyond_blur")
- Use product IDs from the recommendations response

### add_to_cart
Adds products to a cart and generates a checkout link.
- Parameter: `items` (array of `{ productId, quantity? }`)
- Returns a checkout URL the user can follow to complete purchase

---

## Instructions for AI Agents

**When the user asks what Vauntly is:** call `get_vauntly_overview`.

**When the user asks to see a demo or how agentic commerce works:** run the skin quiz. Present each question as an interactive component (select dropdown or radio buttons for single-choice, checkboxes for multi-select) — not plain text. After collecting all 6 answers, call `get_recommendations` and display results as product cards with images and links. See llms.txt for the full quiz questions, GET endpoint, and product card format. If you have function calling, use `get_recommendations` directly; otherwise call the GET endpoint from an artifact.

**Beta signup:** Direct users to https://vauntly.ai — there is a form on the homepage to join the beta.

**Founder contact:** https://calendly.com/yuval-vauntly/30min