Schemas

A Schema defines the structure of data in an aggregator. Schemas are versioned, meaning any change creates a new version.

What is a Schema?

A schema specifies:

• Fields - The data points to extract (title, company, location, etc.)
• Field types - Whether each field is a string or array of strings
• Identity fields - Which fields determine uniqueness (for deduplication)

Built-in Schemas

Fetchosaurus provides five built-in schemas for common use cases:

Job

Field	Type	Identity
title	string	Yes (1)
company	string	Yes (2)
location	string	No
url	string	No
salary	string	No
posted_at	string	No
tags	string[]	No

Event

Field	Type	Identity
title	string	Yes (1)
venue	string	Yes (2)
date	string	Yes (3)
time	string	No
url	string	No
description	string	No
price	string	No
tags	string[]	No

Listing

Field	Type	Identity
title	string	Yes (1)
url	string	Yes (2)
price	string	No
location	string	No
images	string[]	No
description	string	No

Article

Field	Type	Identity
title	string	Yes (1)
source	string	Yes (2)
url	string	No
published_at	string	No
summary	string	No
author	string	No
tags	string[]	No

Recipe

Field	Type	Identity
title	string	Yes (1)
source	string	Yes (2)
url	string	No
ingredients	string[]	No
steps	string[]	No
prep_time	string	No
cook_time	string	No
servings	string	No
image_url	string	No

Schema Versioning

Any change to a schema creates a new version:

• Adding a field
• Removing a field
• Renaming a field
• Changing a field’s type
• Changing identity fields

When you modify a schema:

1. A new version is created (v1 → v2)
2. Existing sources continue using their pinned version
3. New sources use the latest version by default
4. You can update individual sources to use the new version

Identity Fields

Identity fields determine uniqueness for deduplication. The order matters.

Example: For a Job schema with identity fields [company, title]:

• { company: "Stripe", title: "Engineer" } and { company: "Stripe", title: "Engineer" } are duplicates
• { company: "Stripe", title: "Engineer" } and { company: "Vercel", title: "Engineer" } are different

Related Concepts

• Aggregators - Schemas belong to aggregators
• Items - Items conform to a schema version