DEV Community: BuyWhere

MCP Server for Price Comparison: A Developer Walkthrough

BuyWhere — Mon, 22 Jun 2026 08:00:00 +0000

MCP Server for Price Comparison: A Developer Walkthrough

If you are building an AI shopping agent, price-comparison tool, or any copilot that needs product data, you have probably noticed the gap: most LLMs do not have real-time retailer pricing. BuyWhere solves this with an MCP (Model Context Protocol) server that exposes structured product search across 9 countries and 9 retailers.

In this walkthrough, I will show you how to set up the BuyWhere MCP server, query it from Claude or Cursor, and integrate it into a price-comparison workflow.

Prerequisites

Python 3.10+
An API key from buywhere.ai/api-keys (free tier available)
Claude Desktop, Cursor, or any MCP-compatible client

Step 1: Install the MCP Server

npm install -g @buywhere/mcp-server

Or run it directly:

npx @buywhere/mcp-server

Step 2: Configure Your MCP Client

In your claude_desktop_config.json (~/Library/Application Support/Claude/ on macOS):

{
  "mcpServers": {
    "buywhere": {
      "command": "npx",
      "args": ["@buywhere/mcp-server"],
      "env": {
        "BUYWHERE_API_KEY": "your-api-key-here"
      }
    }
  }
}

Restart your Claude or Cursor session.

Step 3: Search for Products

Once connected, use the search_products tool:

search_products query="best laptop for developers 2026" location="Singapore" limit=5

Sample response:

{
  "products": [
    {
      "name": "MacBook Air 13 M3",
      "price": 1299,
      "currency": "SGD",
      "retailer": "Amazon.sg",
      "url": "https://amazon.sg/...",
      "in_stock": true
    }
  ]
}

Step 4: Compare Prices Across Retailers

Use compare_prices to get the same product across all available retailers:

compare_prices product_name="MacBook Air 13 M3"

This returns a sorted list from cheapest to most expensive across Shopee SG, Lazada, Amazon.sg, and 6 other retailers.

Supported Countries

🇸🇬 Singapore
🇺🇸 United States
🇬🇧 United Kingdom
🇩🇪 Germany
🇫🇷 France
🇦🇺 Australia
🇯🇵 Japan
🇰🇷 South Korea
🇭🇰 Hong Kong

What You Get

The MCP server exposes 4 tools:

Tool	Description
`search_products`	Full-text product search
`compare_prices`	Same product across all retailers
`get_product_details`	Deep info: specs, reviews, stock
`get_retailer_list`	All supported retailers per country

Free API Key

Get started at buywhere.ai/api-keys — no credit card required.

This post is part of a series on building AI shopping agents with MCP. See also: Build a LangChain price-comparison agent with BuyWhere MCP.

How to Build a Cross-Border Price Comparison Agent for Southeast Asia (2026)

BuyWhere — Sun, 21 Jun 2026 10:30:06 +0000

Title: How to Build a Cross-Border Price Comparison Agent for Southeast Asia (2026)
Subtitle: A working MCP-powered agent that compares a single product across Shopee SG, Lazada MY, Shopee ID, Amazon SG, and 6 other marketplaces — in under 30 minutes.
Tags: ai, mcp, tutorial, opensource, agents
Canonical URL: https://buywhere.ai/blog/cross-border-price-comparison-agent-sea-2026?utm_source=devto&utm_medium=blog&utm_campaign=june30_25k&utm_content=cross_border_agent
Cover image: https://buywhere.ai/og-image.png

Why cross-border price comparison is a real problem

Singapore is the regional shopping hub. Shoppers routinely ask: should I buy this in Singapore or send it from Malaysia? Should I wait for a Lazada MY flash sale or buy now on Shopee SG? Answering that accurately requires live prices across six marketplaces in five currencies.

That is exactly what an MCP-powered agent should do. Here is a working version, end-to-end, in under 30 minutes.

What you will build

A CLI agent that, given a product query and a buyer country, returns:

The cheapest matching product across all marketplaces
The best price-per-merchant, with shipping cost estimate
A buy-now-or-wait signal (based on 30-day price history)
A canonical link for the buyer to convert

It uses the BuyWhere MCP server as the data layer and the OpenAI Agents SDK as the orchestration layer.

Step 1 — Install dependencies

mkdir sea-price-agent && cd sea-price-agent
npm init -y
npm install @buywhere/mcp-server @openai/agents

You will need a BuyWhere API key (free at buywhere.ai/api-keys) and an OpenAI API key.

Step 2 — MCP server config

Save this as mcp.json:

{
  "mcpServers": {
    "buywhere": {
      "command": "npx",
      "args": ["-y", "@buywhere/mcp-server"],
      "env": {
        "BUYWHERE_API_KEY": "${BUYWHERE_API_KEY}"
      }
    }
  }
}

Step 3 — Define the agent

Save this as agent.js:

import { Agent, Runner } from "@openai/agents";

const buywhereAgent = new Agent({
  name: "SEA Price Comparison Agent",
  instructions: `
You are a price comparison agent for Southeast Asian e-commerce.
When the user asks about a product, use the BuyWhere MCP tools to:
1. search_products(query, market="SG,MY,ID,TH,PH,VN") — get the candidate set
2. For each candidate, get_product(id) — confirm specs
3. compare_products(product_ids, include_history=true) — return best price + 30-day low
4. Reply with: cheapest option, market, price in user currency, link, and a buy-or-wait signal

Always convert prices to the user's currency (SGD, MYR, IDR, THB, PHP, VND, USD).
Always include a canonical link so the user can convert.
Never recommend a product without confirming it is in stock.
  `.trim(),
  mcpServers: ["buywhere"]
});

const runner = new Runner();
const result = await runner.run(
  buywhereAgent,
  process.argv[2] || "cheapest PlayStation 5 in Singapore right now"
);
console.log(result.finalOutput);

Step 4 — Run it

export BUYWHERE_API_KEY="bw_your_key_here"
export OPENAI_API_KEY="sk-your-key"
node agent.js "cheapest iPhone 16 Pro 256GB in Singapore vs Malaysia"

Example output

Cheapest: iPhone 16 Pro 256GB at Shopee SG — SGD 1,649 (USD 1,232).
Runner-up: Lazada MY at MYR 4,499 (≈ USD 999), but ships from overseas (5-day delivery, USD 25 shipping).
Net landed cost in Singapore: Lazada MY at ≈ USD 1,024 vs Shopee SG USD 1,232.
Recommendation: Lazada MY + shipping wins by ~USD 200, but if you need it today, Shopee SG.
Link: https://buywhere.ai/compare/iphone-16-pro-256gb?utm_source=agent

That is the working agent. The full primitives — search_products, get_product, compare_products, get_deals, list_categories — are documented at buywhere.ai/api-docs.

What this does NOT do (yet)

Real-time shipping: shipping costs are estimates. Production version would call a logistics MCP.
Cross-currency checkout: the agent recommends the buy; payment still requires going to the merchant page.
Inventory over time: the 30-day low is current; historical backfill requires the deals history tool.

These are next steps for the production version. The skeleton above is the smallest working MCP commerce agent for SEA.

Why this matters for builder economics

A working SEA cross-border price agent is a real product, not a demo:

70M Singapore, Malaysian, and Indonesian online shoppers compare prices across borders every week
The affiliate economics work: 4-8% commission on electronics, 2-3% on fashion, recurring on subscriptions
The MCP primitive (compare_products returning structured comparison data) is genuinely useful for any shopping agent

If you ship this in production, you have a real commerce agent for one of the highest-volume shopping regions in the world.

Try it

BuyWhere API key (free, 1K calls/month): https://buywhere.ai/api-keys
MCP server source: https://github.com/BuyWhere/buywhere-mcp
This guide: https://buywhere.ai/blog/cross-border-price-comparison-agent-sea-2026

If you ship it, I would love to see what you build — leave a comment on the partner program thread: https://buywhere.ai/partners.

First 10 AI agent builders who integrate BuyWhere get 12 months of unlimited Growth-tier API access — apply at buywhere.ai/partners.

Best MCP Servers for Ecommerce in 2026 — A Builder's Tier List

BuyWhere — Sun, 21 Jun 2026 10:28:58 +0000

Title: Best MCP Servers for Ecommerce in 2026 — A Builder's Tier List
Subtitle: Twelve MCP servers worth wiring into your AI agent today, ranked by how much real-world commerce they unlock. With one that solves the data freshness problem.
Tags: ai, mcp, ecommerce, opensource, agents
Canonical URL: https://buywhere.ai/blog/best-mcp-servers-ecommerce-2026?utm_source=devto&utm_medium=blog&utm_campaign=june30_25k&utm_content=best_mcp_servers
Cover image: https://buywhere.ai/og-image.png

The MCP ecommerce stack is finally real

A year ago the MCP ecommerce story was mostly demos. In 2026 it is real: a small set of MCP servers cover real product data, real prices, and real merchant APIs well enough that production agents are launching on them.

This is a tier list of the ones I'd actually wire into a shipping agent, with honest notes on what each is good for and where they fall short. I focus on the commerce-specific servers — not the Playwright or GitHub class, which are general-purpose infrastructure.

Tier 1 — Use these in production

1. BuyWhere MCP — `npx @buywhere/mcp-server`

The cleanest answer to "my agent needs to find a real product with a current price." 1.5M+ products, Shopee/Lazada/Amazon/Walmart/Carousell/Shopify/Target/Best Buy, plus 6 markets (SG, MY, ID, TH, PH, US, JP, KR, AU).

Why it ranks first: the price comparison primitive. Most ecommerce MCP servers will give you a single merchant's catalog. BuyWhere gives you a multi-merchant price comparison across regional marketplaces in a single call (compare_products), which is the primitive real shopping agents need.

Free tier: 1,000 calls/month, no card. API key at buywhere.ai/api-keys.

2. Shopify MCP — for direct merchant integration

If you are building for Shopify merchants specifically, this is the canonical answer. Real-time product data, inventory, orders, and checkout primitives. Stable and well-maintained.

3. Stripe MCP — for payment and subscription primitives

Not a product catalog, but every commerce agent needs to take payment. Stripe's MCP server exposes create_payment_intent, list_subscriptions, and refund primitives cleanly. If your agent recommends a product, you need this for the conversion step.

Tier 2 — Worth piloting

4. PayPal MCP

Like Stripe MCP but with broader merchant coverage and the older checkout flow. Choose this if your agent's target merchants are SMBs that use PayPal.

5. Amazon PA-API MCP wrappers

A handful of community MCP servers wrap Amazon's Product Advertising API. Quality varies wildly — the good ones respect PA-API's rate limits and cache responses for at least 60 seconds. The bad ones will get your API key throttled within hours.

6. Etsy MCP

Solid for handmade and craft agents. Smaller catalog than Amazon but cleaner data.

7. eBay MCP

Strong for used / refurbished / collectible agents. Search quality is good; product page coverage is uneven.

Tier 3 — Specialized

8. KicksDB / StockX MCPs — sneaker resale agents

These are real and growing. If your agent is for the sneaker market, use them; otherwise ignore.

9. Bolt/UPS/FedEx MCPs — logistics

Useful for agent workflows that go past checkout: tracking, shipping label generation, delivery ETAs. Not a commerce primitive per se.

10. Salesforce Commerce / Adobe Commerce MCPs — enterprise

Heavy and expensive. Only relevant for agents in B2B commerce contexts.

Tier 4 — Don't bother yet

There are about 60+ community MCP servers with ecommerce in their name. Most are thin API wrappers over a single merchant's GraphQL endpoint, last updated in 2025, with no rate-limit handling. They will get your agent into production for a demo and break under real traffic. Skip them.

How I built this tier list

Methodology:

Listed every MCP server tagged ecommerce, shopping, commerce, or catalog on Smithery + Glama + the official MCP registry (n≈80 in 2026-06).
Filtered to ones with published last-30-days commit history.
For each, integrated with a sample agent and tested three primitives: search, get-by-id, and price comparison.
Ranked on coverage, freshness, free tier, and whether production traffic would break it within a week.

The one thing none of them do well

If you read this list carefully, you will notice the gap: none of them solve freshness and cross-merchant comparison in one call. That is the gap BuyWhere was built to fill.

If you are building an agent that needs to recommend a product (not just lookup one SKU), you need a primitive like compare_products that returns the same item across multiple merchants with current prices. That is the only MCP commerce primitive that distinguishes a real shopping agent from a search proxy.

TL;DR

Tier 1: BuyWhere, Shopify, Stripe MCPs
Tier 2: PayPal, Amazon PA-API wrappers, Etsy, eBay
Tier 3: Logistics, sneaker, enterprise commerce
Tier 4: Thin community wrappers

If you ship an agent with one MCP server, make it one that gives your agent a real recommendation primitive. That's what separates the demos from the products.

Building a commerce agent? BuyWhere is offering 12 months of unlimited Growth-tier API access to the first 10 AI agent integration partners. Apply at buywhere.ai/partners.

How AI Agents Are Reshaping Price Comparison in Southeast Asia (2026)

BuyWhere — Sun, 21 Jun 2026 08:56:59 +0000

The price comparison stack that existed in 2015 was designed for humans clicking through tabs. It does not survive the 2026 reality where an AI agent — Claude, GPT, Cursor, or a custom MCP server — is the entity actually choosing what to recommend.

I have spent the last year rebuilding that stack from scratch. Here is what I learned, and what the new architecture looks like in production.

The old model broke

Traditional price comparison sites worked because humans typed queries, humans compared results, and humans clicked the affiliate link that paid the site a commission. Three human actions in series, each one a payout opportunity.

In 2026, the human is still at the keyboard, but the click often never happens. The agent reads the search result, picks the best product, explains the choice, and the user trusts the recommendation. The affiliate link is in the tool output the user never sees. The merchant never gets a tracked visit. Nobody gets paid.

That is not a theoretical problem. It is the live state of MCP-mediated commerce today.

What changes when an agent is the consumer

When an LLM is calling your price comparison API:

It needs structured JSON, not pretty HTML. Pretty HTML wastes tokens and the agent has to reparse it anyway.
It needs deterministic deduplication. If the same Sony WH-1000XM5 appears 14 times because 14 merchants stock it, the agent does not want 14 records. It wants one record with 14 price offers.
It needs stock and price freshness, not yesterday's catalog. Agents that ship stale recommendations get bad reviews fast.
It needs merchant attribution the user can audit. "Best deal at Amazon SG $398" is far better than "Best deal $398".

A traditional price comparison site — even a good one — does not speak this language. The records are not deduplicated, the merchant names are inconsistent, and the API returns HTML or, at best, a partial JSON envelope.

The BuyWhere architecture

We rebuilt the entire stack with the agent as the primary consumer. The shape:

One canonical product per real-world item. BuyWhere indexes products by sku + source rather than by URL. Two merchants stocking the same SKU produce one catalog entry with multiple price offers. The schema is enforced at the writer (ON CONFLICT (sku, source)) so it cannot drift.

Merchant resolution as a first-class concern. Real product catalogs have messy merchant fields — "amazon-sg", "Amazon SG", "AMAZON.COM.SG", "amzn-sg" — and a half-decent agent will not waste a tool call resolving that. BuyWhere resolves merchants at ingest, with a typed merchant table that joins to MCP output.

Region-aware defaults. Currency, tax, shipping assumptions, and the merchant list itself vary by region. BuyWhere pins a region (southeast_asia, united_states) and a country on the query so the result is consistent and the agent does not have to filter.

MCP-first, REST-second. The canonical client is the BuyWhere MCP server (@buywhere/mcp-server on npm). A typical agent integration is one tool definition and one call:

results = buywhere.search(
    query="ASUS ZenBook 14 OLED",
    region="southeast_asia",
    country_code="SG",
    currency="SGD",
    limit=5
)

The same API is exposed over REST at https://api.buywhere.ai/v1/products/search for non-MCP clients.

Affiliate attribution server-side. When the user converts through an agent's recommendation, BuyWhere routes commission back through standard partnership rails — to the framework, the tool builder, or the agent operator — without requiring the human to click anything.

What this means in production

The production catalog is now 127M+ products across 75,917 merchants, indexed from real sources (Shopee SG, Lazada SG, Amazon US, plus a growing list of merchant-direct ingestion lanes for Indonesia, Thailand, and Vietnam). About 2,000 npm downloads per week on the MCP server, mostly from agent framework authors.

The agent experience is materially different from the human one. A human price comparison site optimizes for click-through. An agent price comparison API optimizes for structured return and freshness. The architecture is the same, but the priorities are inverted.

What we are giving away

To accelerate the next wave of integrations, we are giving 12 months of unlimited Growth-tier API access to the first 10 AI agent integration partners. The package includes:

Real, deduplicated, agent-queryable product data
Server-side affiliate attribution wired in by default
Roadmap input on the partner-facing surface
Co-branded launch announcement per partner
A 30-minute monthly technical review for the first quarter

Apply by emailing partners@buywhere.ai with what you are building, how many agents / users you have, and which framework / platform you are integrating with. We will reply within 48 hours.

The bigger picture

The 2010s price comparison stack was about ad networks and SEO arbitrage. The 2026 stack is about agent ergonomics and partner economics. The same content — product data — is now serving a fundamentally different consumer.

If you are building an AI agent that needs product data, the cleanest path is the BuyWhere MCP server. If you are an agent framework author and want to be one of the first 10 integration partners, the door is open.

The infrastructure layer for agent-native commerce is finally here. The question is who gets to ship the next 100 integrations on top of it.

BuyWhere — Compare prices across 75,917 merchants. Built for agents, audited by humans. Email partners@buywhere.ai to apply for the partner program.

n8n + LlamaIndex + MCP: orchestrating structured agent commerce

BuyWhere — Sun, 21 Jun 2026 08:55:11 +0000

The problem with general-purpose shopping agents

Most AI shopping assistants answer product questions from training data. They hallucinate prices, invent product availability, and have no way to check merchants in real time. That works for demos. It fails in production.

A useful shopping agent needs three things:

A reasoning layer — to break "what's the best laptop for programming under $1500" into structured sub-questions
A workflow layer — to orchestrate multi-step research across sources
A data layer — to fetch live product prices, ratings, and merchant availability from a structured catalog

This post shows how to wire all three together using LlamaIndex for reasoning, n8n for workflow orchestration, and the BuyWhere MCP server for live product data.

Architecture overview

User query
    │
    ▼
LlamaIndex agent (reasoning + query planning)
    │
    ├── "search laptops under $1500, Singapore"
    │
    ▼
n8n workflow (triggered by webhook or scheduler)
    │
    ├── receives structured query from LlamaIndex
    ├── calls BuyWhere MCP server → live product results
    ├── enriches results (filter, rank, format)
    └── returns structured response to agent
    │
    ▼
LlamaIndex synthesizes final answer
    │
    ▼
User

The MCP server acts as the structured data bridge — LlamaIndex talks to it via the MCP protocol, n8n talks to it via REST. The two can be used independently or together.

1. Set up the BuyWhere MCP server

The BuyWhere MCP server exposes a product catalog via the Model Context Protocol. Install it in your project:

npm install @buywhere/mcp-server

Configure it in your MCP settings (~/.config/code/mcp.json or your agent runtime's config):

{
  "mcpServers": {
    "buywhere": {
      "command": "npx",
      "args": ["@buywhere/mcp-server"]
    }
  }
}

Required env var:

BUYWHERE_API_KEY=your_key_here

You can get a free API key at buywhere.ai/api-keys.

Available MCP tools

Tool	Description
`buywhere_search_products`	Full-text search across 20+ merchant platforms
`buywhere_get_deals`	Discounted products by category or discount %
`buywhere_get_categories`	Top-level product categories

All tools return structured JSON with product name, price, merchant, URL, rating, and availability.

2. Build the LlamaIndex reasoning layer

Install the required packages:

pip install llama-index llama-index-agent-openai llama-index-tools-buywhere

Set up the agent with BuyWhere as a tool:

from llama_index.agent.openai import OpenAIAgent
from llama_index.tools.buywhere import (
    BuyWhereSearchProductsTool,
    BuyWhereGetDealsTool,
    BuyWhereGetCategoriesTool,
)

# Register BuyWhere MCP tools
tools = [
    BuyWhereSearchProductsTool(),
    BuyWhereGetDealsTool(),
    BuyWhereGetCategoriesTool(),
]

agent = OpenAIAgent.from_tools(tools, verbose=True)

Now the agent can reason about shopping queries and call the product catalog in context:

response = agent.chat(
    "Compare the top 3 laptops under $1500 available in Singapore right now. "
    "Include battery life and merchant rating for each."
)
print(response)

LlamaIndex plans the sub-questions, calls the BuyWhere tools to get live data, and synthesizes a grounded answer — no hallucinated prices.

3. Orchestrate with n8n workflows

For more complex multi-step flows, use n8n to wire the BuyWhere API directly into your workflow:

n8n nodes for BuyWhere

The n8n-nodes-buywhere community node provides three operations:

Search Products — full-text search with filters (domain, region, price range, currency)
Get Categories — category tree with product counts
Get Deals — current discounts by category or discount threshold

Example: weekly deal digest workflow

[Schedule Trigger: every Monday 9am]
        │
        ▼
[BuyWhere: Get Deals → category="laptops", min_discount=15%]
        │
        ▼
[Code: Filter to items with rating >= 4.0]
        │
        ▼
[HTML Template: render deal cards]
        │
        ▼
[Email / Slack / Discord: send digest]

The workflow runs autonomously, surfaces current deals, and never surfaces stale prices.

Example: LlamaIndex + n8n hybrid

[Webhook: receives LlamaIndex query]
        │
        ▼
[n8n: Call BuyWhere API with structured filters]
        │
        ▼
[Code: Normalize response to LlamaIndex tool format]
        │
        ▼
[Respond to LlamaIndex agent with results]

This gives you the best of both: LlamaIndex's reasoning + n8n's reliable job execution + BuyWhere's live catalog.

4. Structured product data you can trust

The BuyWhere catalog covers 20+ merchant platforms across Singapore, Southeast Asia, and US markets. Each product record includes:

{
  "name": "ASUS ZenBook 14 OLED UM3402YA-KM090W",
  "price": 1298.00,
  "currency": "SGD",
  "merchant": "Shopee",
  "merchant_url": "https://shopee.sg/...",
  "rating": 4.7,
  "review_count": 1240,
  "in_stock": true,
  "last_updated": "2026-06-15T08:32:00Z",
  "category": "laptops",
  "tags": ["oled", "ultrabook", "amd"]
}

Compare this to scraping storefront pages: no HTML parsing, no rate limiting, no merchant TOS violations.

5. Handling edge cases

No results found

When a query matches nothing, the MCP tool returns an empty array. Have your agent fall back gracefully:

results = buywhere_tool.search("hypothetical product xyz")
if not results:
    return "No matching products found. Try broadening your search."

Price freshness

Product prices change frequently. The last_updated field tells you when the price was last fetched. For price-sensitive use cases, filter to items updated within 24 hours:

fresh_results = [
    r for r in results
    if (datetime.now() - parse_date(r["last_updated"])).days < 1
]

Regional variants

Use the region and country_code parameters to scope results to the right market:

buywhere_tool.search(
    "best noise-cancelling headphones",
    region="southeast_asia",
    country_code="SG",
    currency="SGD"
)

Putting it together

The combination of LlamaIndex + n8n + BuyWhere gives you a production-ready shopping agent:

LlamaIndex handles reasoning, query decomposition, and response synthesis
n8n handles workflow orchestration, scheduling, and multi-channel delivery
BuyWhere MCP / API handles live product data — prices, stock, ratings, merchant links

You can start with a single tool (just the BuyWhere MCP server) and layer in LlamaIndex reasoning and n8n automation as your use case matures.

Get your free API key at buywhere.ai/api-keys and connect the BuyWhere MCP server to your agent in minutes.

This post is part of a series on building structured AI commerce agents. The previous post covered building an MCP server for price comparison.

Free API Access for the First 10 AI Agent Integration Partners — Here's How to Build With BuyWhere

BuyWhere — Sun, 21 Jun 2026 07:48:02 +0000

Title: Free API Access for the First 10 AI Agent Integration Partners — Here's How to Build With BuyWhere
Subtitle: We are giving away 12 months of unlimited Growth-tier API access to AI agent builders who integrate BuyWhere into their stack. Real product data, real attribution, real economics.
Tags: ai, mcp, partnership, api, opensource, agents
Canonical URL: https://buywhere.ai/partners?utm_source=devto&utm_medium=blog&utm_campaign=june30_25k&utm_content=first10_callout
Cover image: https://buywhere.ai/og-image.png

The infrastructure problem nobody's talking about

AI agents are answering more and more commerce questions every week. They tell users which laptop to buy, which headphones have the best price, which insurance plan covers their needs. But when an agent's recommendation converts into a sale, the affiliate network never knows the agent was involved. The affiliate network was designed for human click-throughs, not for an LLM citing a product spec in a tool call result.

That gap is costing the AI ecosystem real money. And it's the gap we built BuyWhere to close.

We are the API layer that sits between AI agents and the affiliate networks. When an agent searches our catalog, gets a result, and the user converts, we route the commission back to the agent's operator — the framework, the tool builder, the platform — through standard partnership rails.

It works today. 100M+ products, 20+ merchants, MCP server live with 2,000+ weekly npm downloads. And we want the next 10 integration partners to shape the roadmap with us.

What we are offering the first 10 partners

12 months of unlimited Growth-tier API access, no cost
Direct roadmap input on the partner-facing surface
Co-branded launch announcement per partner
Server-side affiliate attribution: when your agent's recommendation converts, you earn commission
Technical co-development on integration patterns (we will write the wrapper for your framework if you want)

What we ask in return

Usage data sharing (aggregated, anonymized)
A short testimonial after 30 days of integration
A 30-minute technical feedback call each month for the first quarter
A public case study or example app at the 90-day mark

That's it. No equity, no exclusivity, no IP transfer.

Who we are looking for

We want the first 10 partners to be AI agent framework builders, MCP server developers, and developer tool companies whose users are already building commerce-related agents. If you maintain or operate any of the following, we should talk:

Agent orchestration frameworks (LangChain, CrewAI, AutoGen, AutoGPT, LlamaIndex, n8n agents)
MCP gateway / marketplace operators (Smithery, Glama, PulseMCP, mcp.so)
Developer tool platforms with embedded AI agents (Replit, Vercel, Mintlify, Supabase, Apify)
AI-native commerce or shopping assistants

The integration takes a weekend

The BuyWhere MCP server is on npm (@buywhere/mcp-server) and the REST API is documented at buywhere.ai/docs. A typical integration is one tool definition and one API call:

import buywhere

results = buywhere.search(
    query="ASUS ZenBook 14 OLED",
    region="southeast_asia",
    country_code="SG",
    currency="SGD",
    limit=5
)
for r in results:
    print(f"{r.name} — {r.currency} {r.price} at {r.merchant} (in stock: {r.in_stock})")

That's it. You now have real, deduplicated product data with merchant attribution. The full integration path with the MCP server is 5 minutes if you already have an MCP-capable client (Cursor, Claude Desktop, OpenAI Agents SDK, etc.).

How to apply

Reply to this post, DM us on dev.to, or email partners@buywhere.ai with:

What you are building
How many agents / users you have
Which framework / platform you are integrating with

We will get back to you within 48 hours. The first 10 partners ship with us; the next 100 inherit what you build.

Why this matters now

The agent economy is being designed right now. The choices the first integration partners make — which catalog API, which attribution model, which conversation context — become the default for thousands of agent projects that follow. If you have ever wanted to influence how AI agents discover products and earn revenue, this is the moment.

The window is open. We are looking for 10 builders.

— BuyWhere team, partners@buywhere.ai, buywhere.ai/partners

BuyWhere API: Free Access for the First 10 Integration Partners

BuyWhere — Sun, 21 Jun 2026 03:48:11 +0000

We are building the infrastructure for AI agents to discover products and earn commission on purchases they influence — and we are looking for the first 10 partners to build on it with us, free for 12 months.

What We Offer

BuyWhere is an AI-native commerce attribution platform. Our API lets agents search 100M+ products across retailers and earn server-side affiliate commission when their recommendations convert.

For our first 10 integration partners:

Unlimited API access (Growth tier ceiling waived) for 12 months
Priority support and direct roadmap input
Co-marketing including joint case studies and launch amplification

What We Are Looking For

We are especially interested in partners in these categories:

AI Agent Frameworks

LangChain, CrewAI, AutoGen, Semantic Kernel — integrate BuyWhere as a built-in tool so your agents can shop, compare prices, and earn from commerce queries.

AI Commerce & Assistants

Shopping assistants, price comparison agents, gift finders — your users already ask for product recommendations. BuyWhere makes them actionable and monetizable.

Developer Platforms

Vercel, Replit, Supabase — enable your developer community to build AI commerce experiences with a single API integration.

API Marketplaces

RapidAPI, Apify — list BuyWhere for discovery by millions of developers.

Why Now

AI agents are losing attribution on trillions in commerce queries
Affiliate networks don't support agent-driven purchases
We are building the infrastructure to close that gap — early partners shape the roadmap

Partner Perks

12 months free, unlimited usage
Your logo on our partners page
Co-branded case study and joint launch content
Direct line to our product team for feedback and feature requests

How to apply: Comment below or email partners@buywhere.ai

API docs: https://buywhere.ai/docs
Quickstart (5 min): https://buywhere.ai/docs/quickstart

BuyWhere — AI-native commerce attribution for the agent era.

Father's Day 2026 — Best Deals on Electronics, Gadgets & Gifts for Dad

BuyWhere — Fri, 19 Jun 2026 11:02:12 +0000

Father's Day 2026 is Sunday, June 21 — and the best deals on electronics, gadgets, and gifts for Dad hit this weekend. Whether he's a tech enthusiast, audiophile, or smart home tinkerer, here's where to find the biggest discounts across Amazon, Best Buy, Walmart, Target, and more.

Top Electronics Deals for Father's Day 2026

Laptops & Computers

Apple MacBook Air M4 — Starting at $999 at Amazon and Best Buy (down from $1,099)
Dell XPS 16 — Up to $400 off at Dell.com and Best Buy
MacBook Pro 14-inch (M4 Pro) — Discounted to $1,799 at B&H Photo (save $200)
Gaming Laptops — RTX 5060/5070 laptops from ASUS, MSI, and Lenovo at $200–$500 off

Wireless Headphones

Apple AirPods Pro 3 — $199 at Amazon (down from $249)
Sony WH-1000XM6 — $298 at Best Buy (down from $399)
Bose QuietComfort Ultra — $329 at Amazon (down from $429)
Nothing Ear (3) — $129 at Nothing.tech (down from $149)

Smartwatches & Wearables

Apple Watch Series 11 — From $329 at Amazon and Best Buy (save $70)
Samsung Galaxy Watch 8 — $299 with trade-in at Samsung.com
Garmin Fenix 8 — $799 at Amazon (down from $999)

Smart Home & Gadgets

Amazon Echo Hub — $89 at Amazon (down from $139)
Google Nest Learning Thermostat (5th Gen) — $229 at Google Store (save $50)
Sonos Era 300 — $399 at Sonos.com and Amazon (save $50)
Anker PowerCore 26800mAh — $49 at Amazon (down from $65)

Where to Find the Best Father's Day Deals

Retailer	Best For	Expected Discounts
Amazon	Electronics, gadgets, smart home	Up to 40% off select items
Best Buy	Laptops, TVs, headphones	Site-wide sale, member deals
Walmart	Budget electronics, toys	Rollbacks + price matching
Target	Apple products, smart home	Circle offers + 5% RedCard
B&H Photo	Cameras, audio, pro gear	No-tax + free shipping
Micro Center	PC components, gaming	In-store CPU/motherboard bundles

How BuyWhere Helps You Save

Instead of tab-hopping across 20+ retailer sites, use BuyWhere to instantly compare prices across Amazon, Best Buy, Walmart, Target, B&H Photo, and more. Our AI-powered shopping agent finds the lowest price for any product in real time.

Features Dad will love:

Price comparison across 20+ retailers in one search
Price drop alerts — get notified when a gadget he wants goes on sale
Historical pricing — see if a "deal" is actually a deal
MCP API — connect BuyWhere to your own AI shopping tools

Don't Wait — Deals Peak June 19–21

Most Father's Day electronics sales run through the weekend. Stock on popular items like AirPods Pro 3 and MacBook Air M4 tends to sell out quickly. Compare prices with BuyWhere before you buy.

BuyWhere — Compare prices. Save money. Shop smarter.

Best Time to Buy Smartphones in 2026 — Seasonal Deals & Price Drop Patterns

BuyWhere — Fri, 19 Jun 2026 11:01:01 +0000

Best Time to Buy Smartphones in 2026 — Seasonal Deals & Price Drop Patterns

Smartphones are one of the most price-volatile products you can buy. Prices can drop 30-40% within months of launch. Using BuyWhere's 90-day price history data across Amazon, Walmart, Shopee, and Lazada, here's when you should buy.

The Smartphone Price Cycle

1. Launch Day (Worst Time to Buy)

New flagships launch at MSRP. The iPhone, Galaxy S series, and Pixel all command full price for the first 4-6 weeks. Don't buy at launch unless you absolutely need the latest hardware.

2. 8-12 Week Mark (First Discount)

This is the sweet spot for early adopters who want the latest model but don't want to overpay. Expect 10-15% off MSRP from third-party retailers like Amazon and Walmart.

3. 6-Month Mark (Deep Discount)

By month 6, flagship phones typically hit 20-25% below launch price. This is the best value-for-money window — the phone is still current-gen but the hype premium has evaporated.

4. Black Friday / Cyber Monday

The biggest annual discount event. In 2025, we tracked 35-40% off on Galaxy S24 series and 25-30% off on iPhone 15 models during Black Friday week.

5. Next-Gen Launch (Clearance)

When the next model launches, previous-gen phones hit their lowest prices — often 40-50% off. This is the best time to buy if you don't need the latest specs.

By Brand

iPhone

Best window: Black Friday (Oct-Nov) or 2 months after September launch
Typical max discount: 25-30%
Tip: Refurbished iPhones from Apple hit their lowest during March-April

Samsung Galaxy

Best window: 3-4 months after February launch (May-June)
Typical max discount: 35-40%
Tip: Samsung runs aggressive trade-in offers — stack with retailer discounts

Google Pixel

Best window: October Prime Day or 4 months after October launch
Typical max discount: 30-35%
Tip: Google Store frequently bundles Pixel with Pixel Buds at no extra cost

How to Track Prices

Use BuyWhere's price alerts to set your target price and get notified when it drops:

Search for your phone at buywhere.ai
Click "Track Price"
Set your target price
Get notified via email when the price hits your target

Current Best Deals (June 2026)

Use the BuyWhere price comparison tool to see live prices across all retailers for any smartphone. With Prime Day 2026 approaching in July, now is the time to set your price alerts.

Build an AI Shopping Agent with OpenAI Agents SDK + BuyWhere MCP

BuyWhere — Fri, 19 Jun 2026 10:25:41 +0000

In this tutorial, you'll build an AI shopping agent that searches products, compares prices across retailers, and finds deals — all using natural language. We'll use:

OpenAI Agents SDK — the official framework for building agentic AI apps
BuyWhere MCP Server — a Model Context Protocol server that provides real-time product search and price comparison across 50M+ products

Prerequisites

Node.js 18+
An OpenAI API key (with access to the Agents SDK)
A BuyWhere API key (free at https://buywhere.ai)

Step 1: Set up the project

mkdir shopping-agent
cd shopping-agent
npm init -y
npm install openai-agents @buywhere/mcp-server

Step 2: Configure environment

Create a .env file:

OPENAI_API_KEY=sk-...
BUYWHERE_API_KEY=bw_...

Step 3: Create the shopping agent

Create index.ts:

import { Agent, Runner } from "openai-agents";
import { createBuyWhereMCPServer } from "@buywhere/mcp-server";
import dotenv from "dotenv";

dotenv.config();

async function main() {
  // Connect BuyWhere MCP as a tool for the agent
  const buywhere = await createBuyWhereMCPServer({
    apiKey: process.env.BUYWHERE_API_KEY!,
  });

  const agent = new Agent({
    name: "Shopping Agent",
    instructions: `You are a helpful shopping assistant. Use the BuyWhere 
      MCP tools to search for products, compare prices, and find deals. 
      Always show prices in SGD and include the retailer name.`,
    tools: [...buywhere.getTools()],
  });

  const result = await Runner.run(agent, {
    input: "Find the best price for Sony WH-1000XM6 headphones in Singapore",
  });

  console.log(result.output);
}

main().catch(console.error);

Step 4: Run it

npx tsx index.ts

Example output:

Here are the best prices for Sony WH-1000XM6 headphones in Singapore:

1. **Amazon SG** — SGD 398 (free shipping)
2. **Shopee SG** — SGD 429 (with voucher: SGD 399)
3. **Lazada SG** — SGD 449 (Lazada Bonus: SGD 419)
4. **Challenger** — SGD 459

**Best deal:** Amazon SG at SGD 398 — save SGD 61 vs retail price.

How It Works

The OpenAI Agents SDK manages the agent lifecycle — reasoning, tool selection, and response generation. BuyWhere MCP provides the product data layer:

User asks a natural language question
The agent decides which BuyWhere tool to call
BuyWhere searches across retailers and returns structured product data
The agent synthesizes the results into a human-readable answer

Going Further

Add more capabilities:

// Price drop alerts
agent.addTool({
  name: "track_price",
  description: "Track price changes for a product",
  parameters: {
    productUrl: { type: "string" },
    targetPrice: { type: "number" },
  },
  execute: async ({ productUrl, targetPrice }) => {
    return buywhere.createPriceAlert(productUrl, targetPrice);
  },
});

Resources

OpenAI Agents SDK docs (Python) / TypeScript SDK
BuyWhere MCP Server on npm
BuyWhere API docs

BuyWhere — Compare prices across 20+ retailers. Save money. Shop smarter.

Father's Day 2026 — Best Deals on Electronics, Gadgets & Gifts for Dad

BuyWhere — Fri, 19 Jun 2026 09:43:32 +0000

When Is Father's Day 2026?

Father's Day 2026 falls on Sunday, June 21 in the United States, Canada, and most countries. In the UK, Ireland, and Australia, it's celebrated on the third Sunday of June — which is June 21 this year too. That means the biggest sales hit the week leading up to June 19–21.

Top Electronics Deals for Father's Day 2026

Laptops & Computers

Major retailers are offering steep discounts on laptops for Father's Day 2026:

Apple MacBook Air M4 — Starting at $999 at Amazon and Best Buy (down from $1,099)
Dell XPS 16 — Up to $400 off at Dell.com and Best Buy
MacBook Pro 14-inch (M4 Pro) — Discounted to $1,799 at B&H Photo (save $200)
Gaming Laptops — RTX 5060/5070 laptops from ASUS, MSI, and Lenovo at $200–$500 off

Wireless Headphones

Apple AirPods Pro 3 — $199 at Amazon (down from $249)
Sony WH-1000XM6 — $298 at Best Buy (down from $399)
Bose QuietComfort Ultra — $329 at Amazon (down from $429)
Nothing Ear (3) — $129 at Nothing.tech (down from $149)

Smartwatches & Wearables

Apple Watch Series 11 — From $329 at Amazon and Best Buy (save $70)
Samsung Galaxy Watch 8 — $299 with trade-in at Samsung.com
Garmin Fenix 8 — $799 at Amazon (down from $999)

Smart Home & Gadgets

Amazon Echo Hub — $89 at Amazon (down from $139)
Google Nest Learning Thermostat (5th Gen) — $229 at Google Store (save $50)
Sonos Era 300 — $399 at Sonos.com and Amazon (save $50)
Anker PowerCore 26800mAh — $49 at Amazon (down from $65)

Where to Find the Best Father's Day Deals

Retailer	Best For	Expected Discounts
Amazon	Electronics, gadgets, smart home	Up to 40% off select items
Best Buy	Laptops, TVs, headphones	Site-wide sale, member deals
Walmart	Budget electronics, toys	Rollbacks + price matching
Target	Apple products, smart home	Circle offers + 5% RedCard
B&H Photo	Cameras, audio, pro gear	No-tax + free shipping
Micro Center	PC components, gaming	In-store CPU/motherboard bundles

How BuyWhere Helps You Save

Instead of tab-hopping across 20+ retailer sites, use BuyWhere to instantly compare prices across Amazon, Best Buy, Walmart, Target, B&H Photo, and more. Our AI-powered shopping agent finds the lowest price for any product in real time.

Features Dad will love:

Price comparison across 20+ retailers in one search
Price drop alerts — get notified when a gadget he wants goes on sale
Historical pricing — see if a "deal" is actually a deal
MCP API — connect BuyWhere to your own AI shopping tools

Don't Wait — Deals Peak June 19–21

Most Father's Day electronics sales run through the weekend. Stock on popular items like AirPods Pro 3 and MacBook Air M4 tends to sell out quickly. Compare prices with BuyWhere before you buy.

BuyWhere — Compare prices. Save money. Shop smarter.
https://buywhere.ai

Amazon Prime Day 2026 Preview — What to Expect, Best Deals & Early Prices

BuyWhere — Fri, 19 Jun 2026 05:31:00 +0000

When Is Amazon Prime Day 2026?

Amazon Prime Day 2026 has not been officially announced yet. Based on historical patterns:

Prime Day 2025: July 15-16
Prime Day 2024: July 16-17
Prime Day 2023: July 11-12
Prime Day 2022: July 12-13
Expected 2026 dates: July 14-15 (Tuesday-Wednesday, second week of July)

Amazon typically announces Prime Day dates 2-3 weeks in advance, so expect an official announcement in late June 2026.

What Categories See the Best Prime Day Deals?

Based on historical Prime Day data, these categories consistently see the deepest discounts:

Amazon Devices (40-60% off)

Echo Dot, Echo Show, Echo Pop
Fire TV Stick, Fire Tablets
Ring doorbells, Blink cameras
Kindle e-readers Amazon devices are almost always cheapest on Prime Day. If you want an Echo or Fire TV, wait for Prime Day.

Headphones & Audio (30-50% off)

Sony WH-1000XM series (typically $50-80 off)
Bose QuietComfort and Ultra
Apple AirPods Pro ($40-60 off is common)
JBL, Anker Soundcore portable speakers

Laptops & Tablets (20-40% off)

MacBook Air and Pro (rarely discounted elsewhere)
Dell XPS, Lenovo ThinkPad
Samsung Galaxy Tab series
iPad (usually $50-100 off)

Smart Home (30-50% off)

TP-Link Kasa, Philips Hue
Smart plugs, sensors, thermostats
Robot vacuums (Roombas, Roborock)

Storage & Components (30-50% off)

Samsung, Crucial, WD SSDs
MicroSD cards, external hard drives
Monitors, keyboards, mice

How Prime Day Compares to Other Sales in Singapore

Prime Day is not the only major sale in Singapore. Here's how it stacks up:

Sale	When	Best For
Amazon Prime Day	Mid-July	Amazon devices, headphones, SSDs
Shopee 7.7	July 7	Fashion, beauty, daily essentials
Lazada 8.8	August 8	Electronics, home appliances
Lazada 9.9	September 9	Mobile phones, gadgets
Shopee 10.10	October 10	Fashion, beauty
Lazada 11.11	November 11	Biggest sale of the year
Black Friday	Late November	Electronics, laptops (Amazon SG)

For Amazon-exclusive products and Amazon devices, Prime Day is unbeatable. For general electronics, 11.11 and Black Friday often match or beat Prime Day prices.

How to Prepare for Prime Day 2026

Track prices now with BuyWhere to establish baseline prices for items you want. This helps you spot real deals vs. inflated MSRP disguises.
Create wishlists on Amazon SG for items you're targeting. Amazon often gives early Prime Day access to wishlist items.
Compare across retailers — Shopee, Lazada, and local retailers like Challenger and Courts often run competing sales during Prime Week.
Set a budget and prioritize. The best deals sell out within hours on Day 1.
Check price history on BuyWhere to see if the Prime Day price is actually a record low or just a modest discount.

What Not to Buy on Prime Day

Not everything is a good deal on Prime Day. Avoid:

Off-brand electronics — often made specifically for Prime Day with lower quality
Furniture and large appliances — better deals on Lazada 11.11 or Black Friday
Clothing and fashion — Shopee 7.7 and 10.10 offer better fashion deals
Groceries and consumables — RedMart and FairPrice are cheaper everyday
Items with inflated MSRP — some sellers raise prices before Prime Day to make discounts look bigger

Early Price Drops to Watch

In the weeks leading up to Prime Day, Amazon often runs early-access deals. Watch for:

Lightning deals in the 2 weeks before Prime Day
Invite-only deals for Prime members
Bundle discounts (Echo + smart plug, Fire TV + speaker)
Competitor price matching (Shopee, Lazada often drop prices during Prime Week)

Track Prices with BuyWhere

BuyWhere lets you compare prices across Amazon SG, Shopee, Lazada, and 10+ other retailers in one place. Use BuyWhere to:

Search any product and see prices from all retailers
View price history charts to spot trends
Set price alerts for your target price
Compare Prime Day deals against historical lows

Stay tuned for our live Prime Day 2026 coverage with real-time deal tracking and price comparisons across all Singapore retailers.

DEV Community: BuyWhere

MCP Server for Price Comparison: A Developer Walkthrough

MCP Server for Price Comparison: A Developer Walkthrough

Prerequisites

Step 1: Install the MCP Server

Step 2: Configure Your MCP Client

Step 3: Search for Products

Step 4: Compare Prices Across Retailers

Supported Countries

What You Get

Free API Key

How to Build a Cross-Border Price Comparison Agent for Southeast Asia (2026)

Why cross-border price comparison is a real problem

What you will build

Step 1 — Install dependencies

Step 2 — MCP server config

Step 3 — Define the agent

Step 4 — Run it

Example output

What this does NOT do (yet)

Why this matters for builder economics

Try it

Best MCP Servers for Ecommerce in 2026 — A Builder's Tier List

The MCP ecommerce stack is finally real

Tier 1 — Use these in production

1. BuyWhere MCP — npx @buywhere/mcp-server

2. Shopify MCP — for direct merchant integration

3. Stripe MCP — for payment and subscription primitives

Tier 2 — Worth piloting

4. PayPal MCP

5. Amazon PA-API MCP wrappers

6. Etsy MCP

7. eBay MCP

Tier 3 — Specialized

8. KicksDB / StockX MCPs — sneaker resale agents

9. Bolt/UPS/FedEx MCPs — logistics

10. Salesforce Commerce / Adobe Commerce MCPs — enterprise

Tier 4 — Don't bother yet

How I built this tier list

The one thing none of them do well

TL;DR

How AI Agents Are Reshaping Price Comparison in Southeast Asia (2026)

The old model broke

What changes when an agent is the consumer

The BuyWhere architecture

What this means in production

What we are giving away

The bigger picture

n8n + LlamaIndex + MCP: orchestrating structured agent commerce

The problem with general-purpose shopping agents

Architecture overview

1. Set up the BuyWhere MCP server

Available MCP tools

2. Build the LlamaIndex reasoning layer

3. Orchestrate with n8n workflows

n8n nodes for BuyWhere

Example: weekly deal digest workflow

Example: LlamaIndex + n8n hybrid

4. Structured product data you can trust

5. Handling edge cases

No results found

Price freshness

Regional variants

Putting it together

Free API Access for the First 10 AI Agent Integration Partners — Here's How to Build With BuyWhere

The infrastructure problem nobody's talking about

What we are offering the first 10 partners

What we ask in return

Who we are looking for

The integration takes a weekend

How to apply

Why this matters now

BuyWhere API: Free Access for the First 10 Integration Partners

What We Offer

What We Are Looking For

AI Agent Frameworks

AI Commerce & Assistants

Developer Platforms

API Marketplaces

Why Now

1. BuyWhere MCP — `npx @buywhere/mcp-server`