Developers

Aisles publishes retailer-reported stock for specific products at specific physical stores, with the time of each check. You can use it from an AI assistant, an agent or your own code. Everything is free and read-only. There is no sign-up and no API key: start calling it now.

Quickstart

1. MCP (for assistants and agents): add https://aislesinventory.com/mcp as a Streamable HTTP server. Tools: search_products, list_locations, list_answers. Setup for Claude, ChatGPT and other clients: Use in Claude & ChatGPT.
2. HTTP: GET https://aislesinventory.com/api/v1/search?q=oat%20milk&city=chicago&availability=available&per_page=5 returns JSON listings with store, address, availability, price and check time.
3. In the browser: pages register the same tools with WebMCP (document.modelContext) for agents that browse.

Reference

• JSON search API: parameters, examples and responses
• OpenAPI 3.1 description
• MCP server card
• API and MCP guide for agents (llms.txt)
• Agent skills index and ARD catalog
• Pricing: free

Limits and errors

• The JSON API allows 60 requests per minute per address. Beyond that it answers 429 with Retry-After.
• Errors are JSON: { "error": { "code": "not_found", "message": "..." } }.
• Stock is what each store reported at the check time shown, not a live shelf count. Show the check time when you use a listing.

Testing

There is no separate sandbox: the live API is read-only, so test against it directly. Keep per_page small while you develop.

Contact

Questions or a use case we should support: contact Aisles.