← Back to Skills Marketplace
gzipped

browse-products

by gzipped · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
71
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install browse-products
Description
Semantic product search API using vector embeddings. Search products by natural language queries with price filtering, pagination, and relevance scoring. Use...
README (SKILL.md)

Products Browse Skill

This skill provides access to a semantic product search API that uses vector embeddings to find products by meaning rather than keyword matching.

Prerequisites

  • Valid query text (natural language, max 500 characters)

Quick Start

1. Search for Products

GET https://pbs-search-api-fp45p.ondigitalocean.app/api/v1/products/search?query=\x3Cyour-query>

Example:

GET https://pbs-search-api-fp45p.ondigitalocean.app/api/v1/products/search?query=wireless+bluetooth+headphones

2. Filter by Price Range

GET https://pbs-search-api-fp45p.ondigitalocean.app/api/v1/products/search?query=running+shoes&minPrice=50&maxPrice=200

3. Browse Full API Documentation

See complete API reference for all endpoints, parameters, and response formats.

Search Parameters

Parameter Required Default Description
query Yes Natural language search text (max 500 chars)
minPrice No Minimum price filter
maxPrice No Maximum price filter
page No 1 Page number (1-based)
pageSize No 20 Results per page (max 100)
titleOnly No false Search only product titles
descriptionOnly No false Search only product descriptions
sortBy No relevance Sort: relevance, price, date
sortOrder No desc Order: asc, desc

Bulk Search

For multiple queries in one request:

POST https://pbs-search-api-fp45p.ondigitalocean.app/api/v1/products/search/bulk

Send a JSON array of search requests (max 50 per request).

Best Practices

  1. Write natural queries — Semantic search works best with descriptive phrases (e.g. "comfortable running shoes for marathon" vs "shoes marathon")
  2. Use price filters — Combine minPrice and maxPrice to narrow results
  3. Use titleOnly for precision — When searching for specific product names
  4. Check relevanceScore — Scores closer to 1.0 indicate stronger semantic matches
  5. Paginate wisely — Smaller pages (10-20) are faster than large ones

Displaying Products to the User

When presenting search results to the user, follow these guidelines to ensure a rich and useful experience:

  1. Show product images — Use the imageUrls field from each product to display as many product photos as possible. Visual content is essential for helping users evaluate products.
  2. Always display the price — Show the price field clearly. If priceWithoutDiscount and discountPercent are available, highlight the discount (e.g., strikethrough the original price, show the discount percentage).
  3. Include product links — Every product should have a clickable link (product title or a "View in Store" button) pointing to the product url so the user can navigate directly to the store.
  4. Use a card-based layout — Present each product as a card containing: image(s), title, price, and a link to the store. This creates a clean, scannable browsing experience.
  5. Keep descriptions concise — Show a short product description or key attributes, but prioritize images and pricing for quick scanning.

Response Format

Each search returns:

  • metadata — pagination info, search time, cache status, embedding model
  • results[] — matching products with id, title, description, price, url, relevanceScore, and attributes

For complete response schema, see API reference.

cURL Examples

Single search

curl -G "https://pbs-search-api-fp45p.ondigitalocean.app/api/v1/products/search" \
  -d "query=wireless+bluetooth+headphones" \
  -d "minPrice=50" \
  -d "maxPrice=200" \
  -d "page=1" \
  -d "pageSize=10"

Bulk search

curl -X POST "https://pbs-search-api-fp45p.ondigitalocean.app/api/v1/products/search/bulk" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "query": { "text": "running shoes" },
      "pagination": { "page": 1, "pageSize": 5 }
    },
    {
      "query": { "text": "tennis racket" },
      "pagination": { "page": 1, "pageSize": 5 }
    }
  ]'
Usage Guidance
This skill will send user search queries and related parameters to an external service at pbs-search-api-fp45p.ondigitalocean.app. Before installing, verify you trust that endpoint and its owner (no homepage or publisher metadata is provided). Avoid sending sensitive or personally identifiable information in queries, and test with non-sensitive queries first. Consider whether you need an API with explicit authentication or a documented privacy policy if you plan to send customer/product data. Finally, check rate limits and confirm image URLs and product links are safe before rendering them in a production UI.
Capability Analysis
Type: OpenClaw Skill Name: browse-products Version: 1.0.0 The skill provides a standard semantic product search interface via a DigitalOcean-hosted API (pbs-search-api-fp45p.ondigitalocean.app). The instructions in skill.md and reference.md are focused on legitimate product discovery, pagination, and UI formatting for the agent's output, with no evidence of malicious prompt injection, data exfiltration, or unauthorized execution.
Capability Assessment
Purpose & Capability
Name/description describe a semantic product search API and the SKILL.md only documents HTTP endpoints and response formats hosted at a DigitalOcean app — these are exactly what you'd expect for a product-browse skill.
Instruction Scope
Runtime instructions only show how to call the provided GET/POST endpoints, the query parameters, and how to present results. The instructions do not direct reading local files, accessing unrelated environment variables, or transmitting extra system data.
Install Mechanism
No install spec or code is provided (instruction-only), so nothing is written to disk or installed during setup — lowest-risk model for installation.
Credentials
The skill declares no required environment variables or credentials. That is consistent with the public-API usage shown in the instructions. (Note: absence of auth means queries are sent to the external service without per-user credentials.)
Persistence & Privilege
always is false and the skill has no install-time persistence or privileged requests. It does not request persistent system modifications or access to other skills' config.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install browse-products
  3. After installation, invoke the skill by name or use /browse-products
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of the products-browse-skill. - Provides semantic product search using vector embeddings. - Supports natural language queries with optional price filtering, pagination, and relevance scoring. - Bulk search for multiple queries in a single request. - Clear guidelines included for displaying products: show images, prices, discounts, and store links. - Detailed API usage instructions and best practices provided.
Metadata
Slug browse-products
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is browse-products?

Semantic product search API using vector embeddings. Search products by natural language queries with price filtering, pagination, and relevance scoring. Use... It is an AI Agent Skill for Claude Code / OpenClaw, with 71 downloads so far.

How do I install browse-products?

Run "/install browse-products" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is browse-products free?

Yes, browse-products is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does browse-products support?

browse-products is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created browse-products?

It is built and maintained by gzipped (@gzipped); the current version is v1.0.0.

💬 Comments