> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ones1ght.com/llms.txt
> Use this file to discover all available pages before exploring further.

# File Connector (CSV · JSON · Excel)

> Upload CSV, JSON, or Excel files and register them as a data source.

## Overview

The file connector registers CSV, JSON, or Excel files as a data source.
CSV and JSON are parsed directly in the browser, and Excel workbooks are parsed as well.
Direct upload supports files **up to 10MB**, and the original file is kept as the stored source on registration.

> [!NOTE]
>
> This screen is available to **Manager and above** (owner, manager) only. Menu visibility `manage`.

## Steps by format

The file wizard steps change dynamically by format.

**CSV**

Data view · filter → Field mapping → Register

**Excel**

**Sheet · region** → Data view · filter → Field mapping → Register

Select a sheet, then drag from the start cell to the end cell to define the data region (a rectangle). If you don't, the whole sheet is used.

**JSON**

**Select key** → Data view · filter → Field mapping → Register

Drill into the `json_path` to choose the target location. Choosing an **array** moves to the data view; choosing a **single object** maps it as one record.

## Data view · filter

**1. Upload file**

Drag or click to select a file (single file, up to 10MB).

**2. Select header row**

Pick the header row that becomes the column names, from among the first 6 rows. If there is no header, click **"No header"** and name the columns directly in the table.

**3. Add filters (optional)**

Add filters so only matching rows are used. Multiple conditions are combined with AND.

**4. Edit columns**

Review the top-100-row preview, edit column names, and toggle which columns to include or exclude.

You then complete the **Field mapping** and **Register** steps of the [shared wizard](/en/console/agent/data-sources#the-shared-4-step-wizard).

## Stored source · security

> [!NOTE]
>
> The source file is kept as the stored source, and only the settings (columns, types, filters, mapping) are saved.
> Data is not copied — it is derived by applying these settings at read time.

> [!WARNING]
>
> Editing a data source keeps the original file unchanged; only the settings change.


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