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

# Core Concepts

> Key terms and concepts used across the Datachecks platform.

## Workspace

A workspace is the top-level organizational unit in Datachecks. All data sources, assets, validations, comparisons, and jobs belong to a workspace. You can create multiple workspaces to separate projects, teams, or environments (for example, staging vs production).

Only admin users can manage workspaces and add data sources to them.

## Data Source

A data source is a database connection registered in your workspace. Datachecks supports Snowflake, Databricks, BigQuery, PostgreSQL, MySQL, Oracle, Azure SQL, and Sybase.

You connect a data source once under Settings and then reference it across validations, comparisons, and agent runs.

## Asset

An asset represents a dataset (table or view) within a data source. When you add a data source to a workspace, Datachecks discovers and registers its datasets as assets. Each asset tracks schema, column metadata, and data profile information.

Assets are the primary unit that agents operate on.

## Validation

A validation is a rule applied to an asset to check data quality or integrity. Examples include null checks, uniqueness checks, range checks, and custom SQL assertions.

Validations run on a schedule or on demand. When a validation detects a threshold violation, it triggers an alert.

## Comparison

A comparison checks whether data in a source dataset matches a target dataset. Comparisons are used during migration to verify that data has been moved accurately. You can compare row counts, column values, and schema structure.

A comparison produces a job that runs asynchronously and returns detailed results including differential percentages and row-level diffs.

## Structural Comparison

A structural comparison checks the schema of a source dataset against a target dataset — column names, data types, nullability, and primary keys — without examining the actual data values.

## Collection

A collection is a user-defined group of validations. You use collections to organize related validations together for easier monitoring and bulk operations.

## Job

A job is an asynchronous background task. Running a comparison, an agent analysis, or a batch validation all create jobs. You can monitor job status, view progress, and terminate running jobs from the Jobs view.

## Agent

An agent is an AI-powered workflow that automates a specific migration task. Datachecks has four agents:

* **Asset Discovery** — Natural language queries about your data schema and assets
* **Migration Assessment** — Analysis of migration complexity and readiness
* **Data Validation** — AI-assisted validation configuration
* **Translation** — SQL query translation between database dialects

See the [Agents](/agents/overview) section for full details.

## API Key

An API key authenticates programmatic access to the Datachecks API. You can create and manage API keys under Settings. API keys are scoped to a workspace.