Configuration

Enter the tables and columns you want to be accessible by AI

The config/vectorify.php file is the central mechanism for defining how Laravel models are processed and upserted to Vectorify. Through a declarative configuration file, it handles collection (table) definitions, column mappings, metadata extraction, and tenancy configuration.

Global Settings

The configuration file contains several global settings that affect the entire package behaviour:

Setting	Purpose
api_key	Authentication token for Vectorify
tenancy	Tenancy mode: single , multi:column, multi:domain
collections	Array of collection definitions

Collections

As described in the concepts page, collections define how Laravel models are synchronized with Vectorify and can use three different syntax patterns. Each collection specifies the data source, field mappings, transformations, and metadata definitions.

Simple Model Reference

The simplest collection configuration references a model class:

'collections' => [
    \App\Models\Invoice::class,
]

This approach uses the model’s $fillable or a custom $vectorify property as the column list.

Resource Class Integration

The resource configuration allows using Laravel API Resources for data transformation:

'collections' => [
	'invoices' => [
		'query' => fn () => \App\Models\Invoice::query()->with('customer'),
		'resource' => \App\Http\Resources\InvoiceResource::class,
		'metadata' => [
			'customer_name' => [
				'type' => 'string',
			],
			'status' => [
				'type' => 'enum',
				'options' => ['draft', 'sent', 'paid'],
			],
			'due_date' => [
				'type' => 'datetime',
			],
		],
	],
],

query is a callable which returns Illuminate\Database\Eloquent\Builder.

Named Model Columns

Complex collections use named keys with configuration arrays:

'collections' => [
	'invoices' => [
		'query' => fn () => \App\Models\Invoice::query()->with('customer'),
		'columns' => [
			'customer' => [
				'relationship' => true,
				'columns' => [
					'name' => [
						'alias' => 'customer_name',
						'metadata' => true,
						'type' => 'string',
					],
				],
			],
			'status' => [
				'metadata' => true,
				'type' => 'enum',
				'options' => ['draft', 'sent', 'paid'],
			],
			'amount',
			'currency_code' => [
				'alias' => 'currency',
			],
			'due_at' => [
				'alias' => 'due_date'
				'format' => 'Y-m-d',
				'metadata' => true,
				'type' => 'datetime',
			],
		],
	],
],

query is a callable which returns Illuminate\Database\Eloquent\Builder.

Column Configuration

Complex columns use arrays to specify transformation and metadata options:

Option	Purpose	Example
alias	Rename field in output	'alias' => 'customer_name'
data	Include in data payload	'data' => false
type	Data type for formatting	'type' => 'datetime'
format	Date/time formatting	'format' => 'Y-m-d'
metadata	Include in metadata	'metadata' => true
tenant	Use for multi-tenancy	'tenant' => true
relationship	Define a relationship	'relationship' => true

Relationship Configuration

Relationships are configured using nested column structures:

'customer' => [
    'relationship' => true,
    'foreign_key' => 'customer_id', // optional
    'columns' => [
        'name' => [
            'alias' => 'customer_name',
            'metadata' => true,
        ],
    ],
],

Metadata Configuration

Metadata defines additional filterable attributes that are stored separately from the row data.

Column-Level Metadata

Metadata can be defined at the column level using the metadata flag:

'status' => [
    'metadata' => true,
    'type' => 'enum',
    'options' => ['draft', 'sent', 'paid'],
],

Global Metadata

Alternatively, metadata can be defined globally, especially when using resource class:

'metadata' => [
    'customer_name' => [
		'type' => 'string',
	],
    'status' => [
        'type' => 'enum',
        'options' => ['draft', 'sent', 'paid'],
    ],
    'due_date' => [
		'type' => 'datetime',
	],
],

Metadata Types

Supported metadata types include:

Type	Purpose	Additional Config
string	Text and numeric values	None
datetime	Date/time values	format option
enum	Enumerated values	options array

Assistant

Responses are generated using AI and may contain mistakes.