Neonjelly: Live Shopify Store Intelligence Built for AI Agents

Neonjelly has indexed 1.37 million Shopify stores—each with traffic estimates, modeled revenue signals, pricing data, and in-store app detection covering 21,270 apps per store. The catalog extends to 368.9 million individual products with price history and seller counts. The underlying data layer is powered by EcomScout indexing, which provides the real-time signal layer that makes price tracking and assortment change detection possible rather than just serving historical snapshots. A brief note on "modeled revenue": the figures are statistical signals estimated from publicly observable store behavior—traffic patterns, app stack, product volume—not accounting records. The site labels them explicitly as modeled, which is the right frame: directionally useful for market sizing and prioritization, not source-of-truth for precise revenue forecasting. What Neonjelly changes is the query interface. Because the catalog is exposed as a remote MCP server, your AI tool can retrieve filtered store lists, look up specific competitor products, or run a saturation check on a product idea using natural language—without any data engineering in between. The question "which Shopify stores in the pet niche between $1M–$5M estimated revenue have added a subscription app in the last 90 days?" is answerable from inside a Cursor or Claude session without intermediate tooling.

Installation varies by platform. Cursor and Claude connect via a single deeplink that opens the app and activates the MCP server in one step. ChatGPT, VS Code, and Windsurf require copying a URL or editing a config file—still fast by developer standards, but two to three steps rather than one. Once connected, the workflow is consistent across all five: provide your API key and start querying with no dashboard to configure and no schema to learn before you can make a useful call. The 60 tools are organized around concrete jobs rather than generic API endpoints. Competitor analysis tools retrieve rival stores' product catalogs and pricing. Dropship tools check saturation—how many sellers carry a product—alongside price history and margin signals. Outreach tools return available emails and social handles from stores matching a filter profile. Due diligence tools pull modeled revenue, traffic trends, and app stack for acquisition evaluation. For developers building AI agents that need market intelligence, Neonjelly is callable as a dependency in the agent's tool set. For non-developers running AI assistants, the natural language interface removes the technical gap—describe the research job and the AI orchestrates the right tool calls. Both paths converge on the same catalog.

One-time queries cover most research tasks. Ongoing competitive intelligence requires something different: a way to monitor rivals continuously without running the same query every day. Neonjelly's signals/watches system handles this—but it's a paid-tier feature, starting with the Explorer plan at $29/month. The free trial doesn't include Signals access. Each watch tracks a set of stores or products and surfaces changes—price moves, product additions, assortment shifts—as automated alerts. Explorer allows 8 concurrent watches; Operator scales to 20; Scale to 40. For a brand tracking three to five core competitors, Explorer's 8 watches are enough to maintain real-time visibility across their key product lines without any manual polling. The watch-based model matters for businesses where competitive pricing is a live variable. Dropshipping margins can turn on a competitor cutting price 20% on a shared SKU. A monitor that surfaces that change the same day it happens is more operationally useful than a weekly manual audit, and considerably cheaper than the data vendor solutions typically used to maintain this kind of coverage.

E-commerce outreach at scale typically involves three separate tools: one to find stores, one to extract contacts, and one to manage the campaign. Neonjelly compresses the first two. Given filter criteria—country, niche, revenue range, app stack—the catalog tools produce a structured store list. The outreach tools then return available emails and social handles for those stores. Two practical constraints worth knowing upfront: the outreach tool returns up to 10 results per call, so building a longer list requires iterative calls with varied filter combinations. Contacts returned are business emails and social URLs only—no phone numbers, no personal owner data—which keeps output on the right side of GDPR and CAN-SPAM compliance for agency and B2B use cases. For agencies prospecting e-commerce clients, brands looking for wholesale or partnership opportunities, and app developers targeting specific Shopify segments by installed app or store size, this compresses a workflow that previously required multiple tools into a single query environment. The filtering specificity is what makes it useful at scale: you're generating a list of stores in a specific niche, above a specific revenue threshold, running a specific app stack—not a generic bulk export.

The clearest fit is anyone running structured e-commerce research inside an AI environment. E-commerce entrepreneurs validating product ideas before sourcing, dropshippers tracking competitor pricing and assortment, agencies prospecting Shopify merchants, and developers building AI agents that need live market context are the core use cases the tool was designed around. Explorer at $29/month is the right starting point for independent researchers—2,000 daily calls comfortably covers a typical research session. A promotional code—LAUNCH—gives 50% off the first three months on any monthly plan, valid through October 31, 2026, bringing Explorer to $14.50/month for the initial period. Operator at $79/month makes sense for teams running automated pipelines or monitoring a broader competitor set. Scale at $199/month targets high-frequency automated workflows where 50,000 daily calls is an operational requirement. The 14-day free trial—100 calls per day, no account, no credit card—lets you evaluate actual data quality on your research questions before committing. The trial is one per device and AI client, and Signals/watches aren't included; those unlock with Explorer. Start the trial in whichever AI environment you use most: Cursor and Claude activate via a single deeplink, making them the lowest-friction entry point.