Useful Data Source For Agents Doing Product Price Sdk
RemoteAbout Useful Data Source For Agents Doing Product Price Sdk
A Singapore BuyWhere catalog API and MCP server for agents researching products and prices. It ranks semantic search results with an auditable value score based on vendor price distributions, reliability, and price rank.
Key features
- 01Semantic product search
- 02Auditable value-score breakdowns
- 03Vendor price distribution analysis
- 04Fuzzy product identity resolution
- 05Causal product comparison
- 06Deal anomaly detection
Use cases
- 01Finding products by natural-language description
- 02Comparing prices and value across vendors
- 03Checking why an item received its value score
- 04Spotting unusually priced listings
MCP support
An MCP endpoint exposes five tools for ranked product search, price distribution, product identity resolution, causal comparison, and deal anomaly detection. MCP-compatible clients connect through the streamable HTTP endpoint.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://useful-data-source-for-agents-production.up.railway.app/mcpclaude mcp add --transport http useful-data-source-for-agents-doing-product-price-sdk https://useful-data-source-for-agents-production.up.railway.app/mcp{"mcpServers":{"useful-data-source-for-agents-doing-product-price-sdk":{"type":"http","url":"https://useful-data-source-for-agents-production.up.railway.app/mcp"}}}Open in Cursor{"servers":{"useful-data-source-for-agents-doing-product-price-sdk":{"type":"http","url":"https://useful-data-source-for-agents-production.up.railway.app/mcp"}}}Open in VS Code{"mcpServers":{"useful-data-source-for-agents-doing-product-price-sdk":{"type":"http","url":"https://useful-data-source-for-agents-production.up.railway.app/mcp"}}}{"mcpServers":{"useful-data-source-for-agents-doing-product-price-sdk":{"serverUrl":"https://useful-data-source-for-agents-production.up.railway.app/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEBuyWhere Singapore Value Intelligence API
Semantic product search over Singapore e-commerce (BuyWhere), ranked by an auditable value-score derived from Shannon entropy across vendor price distributions - not just "cheapest first."
Available both as a plain HTTP API and as an MCP server (5 tools) for AI agents.
Base URL
https://useful-data-source-for-agents-production.up.railway.app
Authentication
All business endpoints require an X-API-Key header:
X-API-Key:
/health requires no authentication.
Pricing
Two ways to pay, same endpoints:
- x402 (pay-per-call, USDC on Base) - currently on Base Sepolia testnet, $0.01/call. A request without payment gets
402 Payment Requiredwith the payment details in thepayment-requiredresponse header. - Stripe (metered billing) - for callers provisioned with an API key and a Stripe customer on the account.
Endpoints
GET /search
Semantic product search over the BuyWhere Singapore catalog, ranked by composite value-score.
Query params:
query(required) - natural-language product descriptionlimit(default 10, 1-50)min_value_score(default 0.0, range 0.0-1.0) - filter out products below this score GET /search?query=noise+cancelling+headphones&limit=5&min_value_score=0.5
Response (ProductSearchResponse):
{
"query": "noise cancelling headphones",
"result_count": 5,
"currency": "SGD",
"products": [
{
"product_id": "abc123",
"title": "Sony WH-1000XM5",
"category": "electronics",
"image_url": "...",
"vendor_offers": ["..."],
"price_min_sgd": 389.00,
"price_max_sgd": 429.00,
"price_mean_sgd": 405.30,
"shannon_entropy_bits": 1.24,
"causal_reliability_score": 0.92,
"value_score": 0.847,
"value_score_explanation": "..."
}
],
"computation_ms": 43.2
}
GET /product/{product_id}/value_breakdown
Full auditable decomposition of the value-score formula for one product - entropy component, reliability component, price-rank component, the formula itself, and the recommended vendor/price.
GET /product/{product_id}/price_distribution
Vendor price distribution for one product: entropy (raw and normalized), price spread, coefficient of variation, and a plain-language interpretation of market fragmentation.
GET /health
Liveness probe. Returns { status, upstream_reachable, api_version, timestamp }. No auth required. Not billed (excluded from both Stripe and x402).
MCP tools
Connect an MCP-compatible client (Claude, Cursor, etc.) to the streamable HTTP endpoint at: https://useful-data-source-for-agents-production.up.railway.app/mcp
Exposes 5 tools:
rank_buywhere_products_by_value_score- ranked search withmax_price_sgdandcategory_filterfetch_buywhere_vendor_price_distribution- vendor spread for a knownproduct_idresolve_buywhere_product_identity- disambiguate a partial/fuzzy product name into candidate product IDscompare_buywhere_products_causal_rank- head-to-head comparison of 2-10 products with confounder adjustmentextract_buywhere_deal_anomalies- z-score based detection of mispriced or outlier listings
Value-score: the method
For each product, prices across vendors are treated as a distribution. The composite value_score blends:
- Entropy component - Shannon entropy of the price distribution across vendors (
H, in bits) normalized against the maximum possible entropy for that vendor count. Higher entropy = genuine price competition, not vendors clustered at the same number. - Reliability component - a causal-reliability score per vendor.
- Price-rank component - where this vendor's price sits in the observed range.
Every response includes the raw components (shannon_entropy_bits, entropy_component, reliability_component, price_rank_component) and, on /product/{id}/value_breakdown, the formula string itself - so an agent can cite the reasoning, not just the number.
Limits
/search:limitup to 50 results per call- MCP
compare_buywhere_products_causal_rank: 2-10 product IDs per call
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.