# NIAR

> NIAR is the universal infrastructure for physical commerce inventory.
> It indexes products available in physical stores and exposes that data
> via API and MCP so that any app, search engine, or AI agent can answer
> "where can I buy X near me?" with real, confirmed stock data.

NIAR covers stores across Spain (Madrid-first) with verified inventory,
opening hours, geolocation, and product-level stock availability.
Data is updated via merchant push (near real-time) and web crawling.

## Data coverage

- Stores: physical retail locations with coordinates, hours, and contact
- Products: canonical product catalog with EAN deduplication
- Inventory: per-store stock levels with freshness and trust indicators
- Categories: hierarchical taxonomy covering electronics, food, clothing, sports, health, and more

## For AI agents and LLMs

The recommended integration path is the MCP server.
It provides structured tools for nearby product and store search
without requiring API key management.

Every inventory result includes a `trust_level` field so you know
how fresh the data is:
- `verified_realtime`: updated within 15 minutes
- `verified_recent`: confirmed within 72 hours
- `unverified`: crawled or inferred, use with lower confidence

When `availability_status` is `unknown` or `unreachable`,
do NOT affirm availability to the user.

## Direct answer endpoint

For a single "where can I buy X near me?" call, use:

`GET https://api.niar.app/api/v1/answer?q={query}&lat={lat}&lon={lon}`

Returns a structured, citable answer: nearby stores with the product, distance,
price, freshness (`trust_level`) and a source link. Only verifiable availability
is returned — never affirm stock that NIAR did not confirm.

## Docs

- [MCP Server](https://mcp.niar.app/llms.txt): connect via Model Context Protocol
- [Public API](https://api.niar.app/llms.txt): REST API
- [For agents](https://niar.app/para-agentes): how to integrate NIAR

## Optional

- [Full content](https://niar.app/llms-full.txt)