> ## Documentation Index
> Fetch the complete documentation index at: https://docs.retailgrid.io/llms.txt
> Use this file to discover all available pages before exploring further.

# How Retailgrid handles your data

> What happens to your data inside Retailgrid - the path from upload to export, how it's stored per organization, and how it's refreshed or removed.

This page explains, in plain terms, what happens to your data once it's in Retailgrid: where it comes in, how it's processed, where it lives, and how it's refreshed or removed. It's written for the person deciding whether to put their catalog, sales, and competitor data into Retailgrid.

> [!NOTE]
>
> For a security review, a data processing agreement, or questions about a specific dataset, [contact us](https://www.retailgrid.io/contact). This page describes the data lifecycle; it isn't a substitute for your security or legal review.

## Your data stays in your organization

Everything you load belongs to your **organization** - the workspace your team shares. Grids, datasets, agent runs, and proposed prices live at the organization level, and Retailgrid scopes reads and writes to the organization you're signed in to. Members, API keys, and connected AI assistants act within that same boundary. See [Account vs Organization](https://docs.retailgrid.io/settings/organization.md) for how the two layers relate.

## The data lifecycle

### 1. Ingest

Data enters Retailgrid three ways:

* **CSV / XLSX upload** in the app, for products, sales, competitor prices, and product-store data. See the [data specs](https://docs.retailgrid.io/data-requirements/datasets.md).
* **Public API** - the same datasets through REST endpoints. See the [API introduction](https://docs.retailgrid.io/api-reference/introduction.md).
* **Integrations** - for example [Shopify](https://docs.retailgrid.io/settings/shopify.md).

Each dataset has a required shape. What's mandatory and what's optional is documented per dataset under [Data specs](https://docs.retailgrid.io/data-requirements/datasets.md).

### 2. Validate

On import, Retailgrid checks the rows before they're accepted - required fields, types, and referential links (for example, a transaction's `item_id` must match a product). Rows that don't pass are reported back rather than silently dropped; CSV imports and the Matching API return row-level errors so you can correct and resend. See [Imports vs bulk](https://docs.retailgrid.io/api-reference/imports-vs-bulk.md).

### 3. Transform and model

Accepted data is organized into your **Grids** and used to compute derived values: metrics (margin, CPI, sales aggregates), elasticity and demand models, forecasts, and agent outputs such as proposed prices. These are **derived** from your source data - the originals aren't overwritten. The account-level model settings that shape these calculations live in [Modelling](https://docs.retailgrid.io/settings/modelling.md).

### 4. Store

Retailgrid keeps a few kinds of data, and it helps to tell them apart:

* **Canonical data** - the products, sales, competitor prices, and stores you loaded. This is the record of truth for your organization.
* **Derived outputs** - metrics, forecasts, proposed prices, and classifications computed from the canonical data. These refresh when the inputs change or when you re-run an agent.
* **Temporary / working data** - staging rows during an import and job intermediates, which are not part of your canonical catalog.

All of it is scoped to your organization.

### 5. Consume and export

You read your data back through Grids, dashboards, and agents in the app; through the [public API](https://docs.retailgrid.io/api-reference/introduction.md) for your ERP or BI tools; and through a [connected AI assistant](https://docs.retailgrid.io/settings/ai-connections.md). Reads reflect live data and are scoped to your organization.

### 6. Retain and remove

Your data stays available while your organization is active. Deleting records through the app or API removes them from your catalog, and derived values that depended on them are recomputed. Canceling a subscription does not by itself delete your data. To request deletion of a dataset or of your organization's data, [contact us](https://www.retailgrid.io/contact).

## Related

* [Data specs](https://docs.retailgrid.io/data-requirements/datasets.md) - the shape of each dataset you load
* [Imports vs bulk](https://docs.retailgrid.io/api-reference/imports-vs-bulk.md) - how imports are validated and applied
* [AI connections](https://docs.retailgrid.io/settings/ai-connections.md) - what a connected assistant can read and do
* [Organization](https://docs.retailgrid.io/settings/organization.md) - the workspace your data lives in
* [Usage & limits](https://docs.retailgrid.io/api-reference/usage-limits.md) - how API and assistant usage is counted


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.