Ask your AI assistant what's on the quick-commerce shelf.
A free, read-only Model Context Protocol server over public shelf data from Blinkit, Zepto and Swiggy Instamart in Bengaluru and Delhi NCR: which brands lead a search, how many stores showed an ad, and who owns the sponsored slots. No account, no API key.
Server URL
https://mcp.shelfxray.in/mcp
Streamable HTTP, no authentication, read-only. Polite limit: 60 requests a minute per IP.
Add it in one click
Copy-paste install
Claude Code
claude mcp add --transport http shelfxray https://mcp.shelfxray.in/mcp
Claude (claude.ai and Claude Desktop)
Settings, then Connectors, then Add custom connector. Name it ShelfXray and paste the server URL above. Leave authentication empty.
Cursor (manual)
Add to ~/.cursor/mcp.json:
{
"mcpServers": {
"shelfxray": { "url": "https://mcp.shelfxray.in/mcp" }
}
}
VS Code (manual)
Add to .vscode/mcp.json:
{
"servers": {
"shelfxray": { "type": "http", "url": "https://mcp.shelfxray.in/mcp" }
}
}
Any client that only speaks stdio
{
"mcpServers": {
"shelfxray": { "command": "npx", "args": ["-y", "mcp-remote", "https://mcp.shelfxray.in/mcp"] }
}
}
Tools and a prompt to try for each
| Tool | What it does | Try asking |
|---|---|---|
get_dataset_info | Coverage, read dates, method, limitations and how to cite. | "What does the ShelfXray dataset cover, and how fresh is it?" |
brand_shelf_snapshot | One brand's rank, store coverage and sponsored slots by platform and city. | "How did Cetaphil show up on Blinkit in Delhi NCR?" |
category_benchmark | Ad pressure, results depth and price range for a category. | "Benchmark the coffee category across platforms." |
search_rank_summary | Who led a search keyword, per platform and city. | "Who led the 'protein bar' search on Zepto in Bengaluru?" |
sponsored_slot_leaders | Brands ranked by sponsored placements owned. | "Which baby care brands own the most sponsored slots in Bengaluru?" |
Test it from a terminal
curl -s https://mcp.shelfxray.in/mcp \
-H 'content-type: application/json' -H 'accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Data, licence and citation
Built on the public Dark Store Index open dataset (CSV and JSON), licensed CC BY 4.0. Please cite: Source: The Dark Store Index, 23 July 2026 – 1 August 2026. https://darkstoreindex.com/data
Only public, shopper-facing shelf reads are served. No brand login, seller account or client data is involved, and nothing is modelled. Method: darkstoreindex.com/about.
Want your own brand read live, store by store, with the rivals who bought the slot above you? Run the free ShelfXray report.