Vector Search in Seconds

Vector search is now available on the Agentic Commerce platform, enabling small and medium-sized merchants to prepare their catalogs with embeddings and surface better results to customers through AI agents.

The Challenge

For small and medium-sized merchants, product discovery is everything. Customers need to find what they're looking for quickly, and when they can't, they leave. Traditional search relies on exact keyword matches—if a customer searches for "comfortable running shoes" but your product is listed as "athletic footwear," they might never find it.

The problem gets worse with AI agents. Autonomous commerce agents need to understand product meaning, not just match keywords. They need to understand that "waterproof jacket" and "rain-resistant outerwear" are the same thing. They need to find products based on what customers are trying to accomplish, not just what words they use.

Vector search solves this by converting product descriptions, attributes, and customer queries into embeddings—mathematical representations that capture meaning. When a customer searches for "something to keep me warm on a hike," the system understands they're looking for outdoor clothing, not just products that contain those exact words.

How It Works

Vector search converts your product catalog into embeddings—dense numerical representations that capture semantic meaning. Each product's description, attributes, and metadata are transformed into vectors that live in a high-dimensional space where similar products cluster together.

When a customer searches, their query is also converted into an embedding. The system then finds products whose embeddings are closest to the query embedding, returning results based on meaning rather than exact text matches.

The platform supports three search modes: vector search for pure semantic matching, semantic search that combines embeddings with relevance scoring, and hybrid search that merges vector results with traditional keyword matching for the best of both approaches.

Why It Matters for Autonomous Commerce

Vector search is integral to successfully creating high-impact autonomous commerce agents. AI agents need to understand product relationships, customer intent, and context—all of which require semantic understanding that goes beyond keyword matching.

When an agent helps a customer find "a gift for someone who loves cooking," it needs to understand that this might mean kitchen tools, cookbooks, or specialty ingredients. Vector search enables this understanding by finding products based on their meaning and relationships, not just their text.

For merchants, this means agents can surface products customers might not have found otherwise, increasing discovery and sales. It means agents can build workflows that understand product context—recommending accessories that match a main purchase, finding alternatives when something is out of stock, or identifying products that complete a set.

Preparing Your Catalog

The platform automatically generates embeddings for your product catalog. You don't need to understand the technical details—the system handles embedding generation, vector storage, and search optimization behind the scenes.

Products are indexed with their descriptions, attributes, categories, and metadata. The system creates embeddings that capture product meaning, style, use cases, and relationships. This happens automatically as you add or update products, ensuring your catalog is always ready for semantic search.

The process is designed for merchants who want results without managing infrastructure. You focus on your products; the platform handles the vectorization, indexing, and search optimization.

Search Capabilities

Vector search supports natural language queries that understand customer intent. A search for "something comfortable for working from home" finds products based on meaning—ergonomic chairs, soft clothing, home office accessories—not just products that contain those words.

The system supports filtering by price, inventory status, attributes, categories, and creation dates. You can boost in-stock items, diversify results to show variety, and set confidence thresholds to ensure quality matches.

Results include semantic confidence scores that indicate how well products match the query's intent. The system can also provide insights about search patterns, price ranges, and aesthetic matches that help merchants understand how customers are discovering their products.

Real-World Example

Here's how a merchant might use vector search to help customers find products for a specific use case:

$ hyperfold search "gift for a creative professional who works from home" \
  --type hybrid \
  --filter "price:<=500" \
  --category "electronics/accessories" \
  --attr "warranty:>=1year" \
  --in-stock-only \
  --boost-in-stock \
  --diversify \
  --min-confidence 0.75 \
  --limit 10

This command searches for products that match the semantic meaning of "gift for a creative professional who works from home" using hybrid search that combines vector and keyword matching. It filters by price, category, and warranty attributes, excludes out-of-stock items, boosts in-stock products in results, diversifies to show variety, sets a minimum confidence threshold, and limits results to the top 10 matches.

The system understands that this query might mean ergonomic equipment, creative tools, home office accessories, or productivity software—and returns products that match the intent, not just the words.

Business Impact

Merchants using vector search see significant improvements in product discovery and customer satisfaction. Customers find products faster because search understands their intent. Agents can surface relevant products even when customers use different words than what's in product descriptions.

The system works at scale, handling thousands of products and millions of search queries. It learns from search patterns, continuously improving its understanding of product relationships and customer intent.

For small and medium-sized merchants, this means competing with larger retailers on product discovery without building custom search infrastructure. It means enabling autonomous commerce agents that understand product context and customer needs.

---

Vector search on the Agentic Commerce platform enables merchants to prepare their catalogs with embeddings and surface better results to customers through AI agents.

Get started to learn how vector search can transform your product discovery.