> ## Documentation Index > Fetch the complete documentation index at: https://docs.mora.com/llms.txt > Use this file to discover all available pages before exploring further. # Core concepts > The building blocks of Mora, datasources, connections, datasets, transformations, charts, and dashboards. Mora has six core concepts that work together. Understanding them will help you get the most out of the platform. ## Datasources A **datasource** is a direct connection to a SQL database or data warehouse. Mora supports seven database types: * PostgreSQL * MySQL * Amazon Redshift * Google BigQuery * ClickHouse * Snowflake * Databricks You provide credentials (host, user, password, etc.) and Mora connects directly to run queries. Each datasource also syncs a **catalog** of schemas, tables, and columns that powers autocomplete and AI assistance. ## Connections A **connection** syncs data from a third-party application into your warehouse datasource. Mora supports connections to: * PostHog, Stripe, Salesforce, HubSpot, Attio, Google Sheets, and Facebook Ads (with setup guides in this documentation). * Hundreds of other apps via [mora.com/integrations](https://mora.com/integrations). Connections use Fivetran under the hood. When you set up a connection, the data lands in a schema inside your existing datasource, where you can query it like any other table. ## Datasets A **dataset** is a saved SQL query tied to a specific datasource. Datasets have a lifecycle: 1. **Draft**: the query you are editing in the SQL editor, edited by you or the AI. Unsaved changes show as a diff against the published query. 2. **Published**: the saved query that powers your charts. Publishing runs the query and materializes the results. Datasets can be **private** (visible only to you) or shared with your organization. ## Transformations **Transformations** are a no-code layer that sits on top of your published dataset results. Instead of rewriting SQL, you can point and click to: * **Group** rows by one or more columns. * **Aggregate** with operations like sum, average, count, median, min, max, and more. * **Filter** with conditions like equals, contains, greater than, between, relative dates, and boolean checks. * **Sort** by any column in ascending or descending order. * **Hide columns** you don't need in the visualization. Transformations are applied in-browser using DuckDB, so they are fast and don't hit your database again. ## Charts (blocks) A **chart** (internally called a "block") is a single visualization that lives on a dashboard. Each chart is linked to a dataset and has one of three display types: | Display | Use case | | -------------- | ------------------------------------------------------------------- | | **Table** | Browse rows, search, and explore raw data. | | **Graph** | Visualize trends and comparisons, line, area, bar, scatter, or pie. | | **Statistics** | Show single-value KPIs with optional period-over-period deltas. | Charts can have their own transformation configuration, so the same dataset can power multiple charts with different groupings, filters, and aggregations. ## Dashboards A **dashboard** is a canvas where you arrange charts on a grid. Dashboards support: * **Drag-and-drop layout** on a 12-column grid. * **Dashboard-level filters** that apply across all charts, linking columns from different datasets. * **Sharing**: dashboards can be favorited, made public within your org, or shared via a copied link. ## AI chat The AI chat panel is available throughout Mora. It can: * Write and debug SQL in the context of your schema. * Create datasets, charts, and dashboards. * Apply transformations and dashboard filters. * Search your catalog to find the right tables and columns. The chat has four modes, **Auto**, **SQL**, **Dashboard**, and **Ask** (read-only), so you can steer it toward the kind of help you need.