diff --git a/ideas/ai-first-ebay-cataloging-assistant/IDEA.md b/ideas/ai-first-ebay-cataloging-assistant/IDEA.md
new file mode 100644
index 0000000..8352000
--- /dev/null
+++ b/ideas/ai-first-ebay-cataloging-assistant/IDEA.md
@@ -0,0 +1,69 @@
+# AI-first eBay Cataloging Assistant
+
+## One-Liner
+
+An AI-powered assistant that turns raw product descriptions or photos into complete, buyer-ready eBay listings.
+
+## Problem
+
+- **Pain point**: Sellers spend significant time researching comparable sold listings, rewriting descriptions into concise paragraphs, selecting attributes, and choosing keywords that balance appeal with policy compliance.
+- **Who experiences it**: Home resellers, shop owners migrating from in-person sales to eBay, and anyone managing a high volume of SKUs.
+- **How painful / frequent**: This is a repeated, time-consuming task that happens for every new or refreshed listing.
+
+## Solution
+
+- **What does it do**: Accepts free-form notes or images, queries and analyzes the market, and outputs formatted eBay listings optimized for discovery and conversion.
+- **Core workflows**:
+  1. Input raw product notes or photos via chat, email, or Shopify-style CSV.
+  2. Research sold and active comps to suggest pricing.
+  3. Generate titles, descriptions, bullet points, item specifics, and keywords.
+  4. Summarize risk factors (too generic, keyword stuffing, suspicious attribute combinations).
+  5. Prepare the listing for review / export.
+- **Minimum Viable Version**:
+  - Text input only.
+  - Research Web + select data sources (eBay sold data or seller guides).
+  - Structured output + copy-ready formatting.
+  - Simple approval workflow.
+
+## Target Market
+
+- **Primary customer segment**: Casual and power eBay resellers.
+- **Market size indicator**: eBay remains one of the largest global marketplaces; resale and small-business growth continue expanding the addressable seller base.
+
+## High Level Use Cases
+
+- Turn a 2-line WhatsApp note into a listing.
+- Bulk-rewrite existing Shopify/Amazon titles into eBay-optimized listings.
+- Suggest realistic prices using sold-listing benchmarks.
+- Enforce listing quality before publish.
+
+## High Level Design / Dependencies
+
+- **Architecture / system design overview**:
+  - Web app + mobile-friendly chat interface.
+  - Multiple research tools in parallel.
+  - Template engine for listing content.
+  - Approval queue with export (eBay file format, JSON, CSV).
+- **External services, data sources, APIs**:
+  - Google Shopping or eBay integration for research.
+  - eBay API for direct publish (later phase).
+  - Vision model for photo descriptions.
+- **Key technical dependencies**:
+  - Hermes agents or equivalent orchestration.
+  - Structured templates + rules for eBay listing quality.
+
+## Differentiation / Competitive Edge
+
+- Chat-native seller workflow (as natural as texting a listing to an assistant) rather than rigid spreadsheets.
+- Compliance-focused generation that avoids policy pitfalls before a human reviewer sees them.
+- Easy integration with product data already captured in other tools or raw notes.
+
+## State of this idea
+
+- `initial`
+
+## Risks
+
+- **Technical**: Compiling reliable, current comparable pricing at scale.
+- **Market**: Seller expectations around publication automation vs. human judgment.
+- **Regulatory**: eBay policy compliance for listing content; data-source licensing.