Why retailerapi exists (and what we deliberately don't try to do)

I've been running ecommerce arbitrage and tooling businesses since 2017. Every year I run into the same wall: there's no clean API for non-Amazon retailer data. Keepa exists for Amazon. CamelCamelCamel exists for Amazon. Honey, Capital One Shopping, every browser extension, every "best deals" affiliate site, all built around Amazon. The other 50% of US ecommerce gets nothing comparable.

Walmart Marketplace has 100M+ SKUs and 240,000+ third-party sellers. Target has 50M+ SKUs. Best Buy, Lowe's, and Home Depot each carry 1M+ SKUs and run on volatile pricing. None of them publishes a price-history API. None of them has a Keepa equivalent.

retailerapi is the tool I wished existed. This post is the case for it, what we're optimizing for, and the things we're explicitly choosing not to do.

The gap

Run this experiment in 2026:

1. Pick any product on Walmart with a UPC. Try to find its 90-day price history. Walmart shows you the current price and that is it.
2. Same product on Target. Same Best Buy. Same Lowe's. None of them surface history.
3. Now check the same UPC on Keepa. If the product is also on Amazon, Keepa shows you a beautiful chart going back 5+ years for Amazon. For Walmart, Keepa shows shallow history. For Target, Best Buy, Lowe's, Home Depot, Keepa shows nothing.

This is the gap. Walmart sellers, multi-retailer arbitrageurs, brand-monitoring teams, repricers, and AI shopping agents all need the data. Nobody is shipping it.

What we're building

A single API that returns:

• Title, brand, image, category for any UPC, EAN, ISBN, GTIN-14, or walmart_item_id
• Current price + offer count from Walmart (50M+ products live)
• Current price from Amazon, eBay, Target, Best Buy, Lowe's, Home Depot (added on demand from a single Google Shopping search per product)
• 30-day price history (preview tier, public)
• Full price history per retailer (gated behind signup)
• Sellers list per retailer (where applicable)
• Reviews + ratings
• Sales rank + estimated velocity (Walmart today; other retailers as data accumulates)

Surfaces:

• REST API at api.retailerapi.com. Token-billed. JSON responses.
• MCP server (@retailerapi/mcp) for AI assistants (Claude Desktop, Claude Code, Cursor).
• Custom GPT in the GPT Store for ChatGPT users.
• Public web pages at retailerapi.com/p/ indexed by Google + cited by LLMs.
• Browser extension (Phase 2) for distributed data collection and consumer use.

What we're NOT trying to do

This list is more important than the previous one. The fastest way to fail at a focused product is to drift into adjacent categories.

We are not Keepa for Amazon. Keepa is excellent at what they do. We're not going to outcompete them on Amazon depth, and we won't try. Phase 3 ships Amazon coverage at a level that's useful for cross-retailer comparison, not at a level that replaces Keepa as your Amazon-only tracker.

We are not a marketplace seller tool. EC (Ecom Circles) is a separate company that handles Walmart Marketplace seller workflows, repricing, and PO management. retailerapi is the product-data layer underneath; it does not run your business operations.

We are not a coupon site. RetailMeNot, Honey, and the affiliate-coupon ecosystem already exist. We focus on product-data and price-history; coupons are noise.

We are not a deal-aggregator destination. Slickdeals, Brad's Deals, DealNews. They optimize for human-curated daily deal feeds. We optimize for programmatic price-data access. (We will run a social deals account for organic exposure, similar to Keepa's X account, but the deals feed is marketing surface, not the product.)

We are not a B2B-only enterprise data vendor. ShopSavvy ProductCloud and Bright Data sell datasets to corporations under custom contracts. We sell self-serve subscriptions starting at $0 (free tier) and $49 per month (entry paid tier). No sales call required.

We are not building our own crawl infrastructure for every retailer. Where Google Shopping returns the data we need (Tier 0), we use that. Where retailer-specific official APIs exist (Best Buy Open API, eBay Browse API, Walmart Marketplace API), we use those. We only fall back to direct scraping (Tier 1) for retailers without official access. The cheapest path that yields accurate data wins.

We are not a price-prediction tool. "Will this drop in the next 30 days?" predictions are interesting and unreliable. We surface the data. You make the call.

We will not sell shopper data. End-users of our customers' tools never become our customers. Buyer behavior is not our product. Our customer is the developer or seller calling our API.

What we're optimizing for

Developer velocity. From sign-up to first API call in under 90 seconds. Free tier with no card. Real documentation. Open-source MCP server. Open-source TypeScript SDK.

Cost transparency. Token cost per call published in docs. No surprise overages (calls fail with 429 rather than auto-charging). Plan upgrades take effect at next renewal, not retroactively.

AEO + LLM-citation friendliness. Public product pages explicitly served to GPTBot, ClaudeBot, PerplexityBot, Google-Extended. Schema.org markup follows Bradley placement rules. llms.txt and llms-full.txt curated, not autogenerated. Goal: when ChatGPT or Claude or Perplexity answers "what's the price of UPC X" or "is there a Keepa alternative for Walmart," retailerapi is the cited source.

Speed of integration for AI agents. MCP support is first-class. We assume your shopping agent needs reliable product data and shouldn't have to scrape its own.

The philosophical bet

Two product-data tools dominated 2010-2025: Keepa for Amazon depth, and Google Shopping for breadth (with poor history). Between them, every consumer-facing comparison tool got built. Honey, PriceBlink, Capital One Shopping, ShopSavvy, all built on top of one or both.

The bet: 2026-2030 will demand a third primitive. AI agents will increasingly do shopping research on behalf of users. Multi-channel sellers will increasingly compare 6 retailers per SKU instead of 2. The product-data layer for these workflows needs an MCP-native, multi-retailer, developer-self-serve API. That layer is retailerapi.

If we're wrong about that bet, we still have a useful Walmart-and-friends API that competes on depth + price + breadth against Keepa Walmart. That's the floor. The ceiling is being the cited source for every AI shopping agent, the integrated MCP backend for every Claude/Cursor/ChatGPT assistant, and the distributed data collection layer for every multi-channel seller in 2030.

Where we are right now

May 2026: Walmart fully live, cross-retailer enrichment for 5 retailers via Tier 0 SERP scrape, MCP server published, REST API live, public pages indexable. Amazon coverage incoming. Browser extension scoped. Social deals poster scoped.

Sign up free → and run your first lookup.

— Matt