Creating a Catalogue

Catalogs in Requestador are used to store predefined values that can be referenced by actions, validation scripts, and transformation scripts.

A catalog acts as a controlled data source that helps guide AI output toward expected business values instead of allowing completely free-form responses.

Catalogs are especially useful when Requestador workflows need to:

• classify content into predefined categories

• validate output against allowed values

• normalize AI-generated responses

• route requests using controlled business logic

• maintain consistency across actions and use cases

In practical use, catalogs help reduce ambiguity and improve output reliability by giving Requestador a structured set of approved values to work with.

The purpose of a catalog is to provide a reusable set of predefined entries that can be used across multiple Requestador actions.

Typical uses include:

• category lists

• classification values

• controlled vocabularies

• predefined labels

• mapping keys used in validation or transformation logic

Catalogs are particularly valuable when AI responses must stay within business-defined options, such as:

• allowed product categories

• SEO categories

• content groups

• routing values

• normalization targets

A catalog contains:

• a catalog name

• an optional description

• optional language support

• one or more catalog entries

Each catalog entry typically includes:

• a Key

• a Value

• an optional Language

The exact way entries are used depends on the action and script logic.

In many cases:

• the Key represents the canonical business identifier

• the Value represents the label or normalized value used in output

Catalogs can be used in Requestador for several purposes.

1. Output validation

A validation script can check whether the AI response contains only values that exist in a catalog.

Example:

• returned category must exist in the categories catalog

2. Output transformation

A transformation script can map an AI-generated value to a normalized catalog value.

Example:

• AI returns a category label

• transform script maps that label to the catalog-defined business value

3. Controlled classification

An action can use catalog values as part of its prompt and script logic to ensure the model selects from a limited set of approved options.

Example:

• assign one category from a predefined list such as:

  • camping-and-hobby

  • DIY

  • electronics

  • fashion

4. Request routing or business logic

Catalogs can also be used to support internal routing logic or decision-making inside Requestador workflows.

Step 1: Open the Catalogs section

In Requestador, navigate to Catalogs.

This page shows the available catalogs for routing requests and for use in Requestador workflows.

Click Add Catalog to create a new catalog.

Step 2: Enter the basic catalog details

Catalog Name

Enter a clear and descriptive catalog name.

Examples:

• categories

• languages

• brand-groups

• seo-categories

The catalog name is important because it is referenced in validation and transformation scripts.

Description

Optionally add a description explaining the purpose of the catalog.

This is useful when multiple users manage actions and catalogs.

Example:

• Allowed category values for product enrichment workflows

Step 3: Enable language support if needed

If your catalog should contain language-specific values, enable Language support.

Use language support when:

• catalog entries differ by language

• values must be localized

• the same key should exist with different language-specific values

If your catalog only contains global values, language support is usually not necessary.

Step 4: Add catalog entries

Each catalog entry contains:

• Key

• Value

• optionally Language

Examples from a category catalog:

• camping-and-hobby → camping-and-hobby

• DIY → DIY

• electronics → electronics

• fashion → fashion

Add entries one by one using the Add button, or upload them in bulk.

Step 5: Upload entries from CSV or JSON if needed

If you already have a larger predefined list of entries, use Upload entries as CSV/JSON.

This is useful when:

• the catalog contains many values

• the values already exist in another source

• you want to initialize or update the catalog in bulk

Step 6: Save the catalog

Click Save.

The catalog is now available for use in Requestador actions and scripts.

Once a catalog has been created, entries can be managed directly from the catalog editor.

Available actions typically include:

• viewing existing entries

• searching entries by key or value

• adding new entries

• editing existing entries

• deleting entries

• exporting catalog data as CSV

This makes catalogs easy to maintain over time as business values change.

Catalog design should stay simple and predictable.

Use stable keys

Catalog keys should be stable business identifiers that do not change often.

Good examples:

• electronics

• household-kitchen

• pets

Use clear values

Values should represent the final business meaning you want the workflow to use.

In simple cases, key and value may be identical.

Keep naming consistent

Do not mix multiple naming patterns in the same catalog.

For example, avoid mixing:

• homeKitchen

• household-kitchen

• HouseholdKitchen

Choose one standard and keep it consistent.

Avoid duplicate meaning

Each catalog should represent one clear business domain.

For example:

• one catalog for categories

• another catalog for languages

• another catalog for internal routing groups

Do not overload a single catalog with unrelated values.

Use catalogs when:

• output must match predefined values

• business rules require controlled vocabularies

• free-form output is too risky

• multiple actions should share the same value source

• classifications must stay consistent across workflows

• scripts need a reusable reference set

Catalogs are especially important in use cases such as:

• category generation

• SEO classification

• language handling

• normalized routing logic

• controlled product enrichment

• use descriptive and stable catalog names

• keep entries consistent in naming and format

• use catalogs for controlled classification workflows

• reuse catalogs across multiple actions where possible

• review and clean up outdated entries periodically

• align catalog values with validation and transformation logic

• test actions whenever catalog values are changed significantly

Using the wrong catalog name in scripts

If the script references a catalog name that does not exist, validation or transformation will fail.

Example:

• script expects catalogs["categories"]

• actual catalog name is different

Mixing multiple purposes in one catalog

A catalog should represent one clear domain. Do not combine unrelated values into the same catalog.

Inconsistent keys and values

If entry naming is inconsistent, AI outputs and script logic become harder to maintain.

Forgetting language design

If values are language-specific, decide early whether language support is required.

Changing catalog values without retesting actions

Catalog changes can affect validation and transformation logic. Always retest dependent actions after important catalog updates.

Catalogs in Requestador provide a reusable and controlled source of predefined values for workflows, scripts, and classification logic.

They are especially useful for:

• output validation

• output transformation

• controlled classification

• normalization

• consistent business logic across multiple actions

A well-designed catalog improves output reliability, reduces ambiguity, and makes Requestador workflows easier to maintain at scale.