SellChad.com

Natural-language product search for e-commerce, built for the AI era.

What it is

SellChad.com adds semantic, intent-aware product search to online stores through an embeddable widget, SDK, and REST API. It ingests product feeds and crawled content, indexes them with vector search, and re-ranks results with business rules — targeting D2C and mid-market e-commerce (10k–1M SKU) and implementation agencies. It also exposes an LLM Commerce Gateway so AI assistants (ChatGPT, Gemini) can query a store’s catalog directly instead of crawling it.

Highlights

• Natural-language semantic search (query → products) with vector ranking, synonyms, boosting, pinning, and A/B testing
• Multi-tenant architecture with hard per-tenant data isolation (Postgres row-level security + per-tenant vector collections)
• Feed ingest (XML/CSV/JSON) plus a JS-rendering crawler with automated taxonomy and attribute mapping
• LLM Commerce Gateway with /.well-known/llm-commerce.json discovery and an OpenAPI/MCP contract for AI-assistant commerce
• SEO/GEO audit crawler and a self-serve analytics/config dashboard
• Ready-made connectors for PrestaShop, WordPress, and Shoper

Stack

TypeScript monorepo · NestJS API · Next.js admin · PostgreSQL (RLS) + MongoDB · Qdrant + Typesense (hybrid keyword + vector) · OpenAI embeddings · RabbitMQ · Redis · Stripe · Docker Compose + Traefik.

Status: Live / production (sellchad.com).