Malloy Explorer: Visual Query Builder for Malloy

What this is: a tour of the Explorer, Publisher's no-code visual query builder. Every action generates valid Malloy against a governed model, so metrics stay correct even across joins. Open it from the Publisher App on any source — the screenshots below use the bundled storefront package.

1. Introduction

Malloy Explorer is a no-code, visual query builder built on top of the Malloy semantic modeling language. It provides a seamless interface for exploring data, analyzing key business metrics, and building reusable queries — all without writing code. Under the hood, every action in the Explorer generates valid Malloy, ensuring semantic consistency & trust while enabling flexible data exploration.

Designed for data consumers and developers alike, Malloy Explorer brings the power of governed semantic models to a modern, interactive user experience. It empowers users to ask complex questions, segment and visualize results, and drill into the details with nested structures — all from a single interface.

Key Benefits

• Semantic-first foundation – Every query runs against a Malloy model, which captures the business meaning and relationships of the data.
• No-code UX – Construct queries visually using dimensions, measures, filters, and charts.
• Live, nested analysis – Easily drill down into rich multi-level breakdowns using Malloy's native nesting capabilities.
• Full transparency – View the Malloy and SQL generated by your visual query.
• AI and developer ready – All queries are valid Malloy code, making them composable, reusable, and LLM-consumable.
• Guaranteed correct metrics - Malloy supports symmetric aggregates, ensuring that metrics compute correctly even across joins with varying cardinality. This eliminates common errors found in traditional SQL-based tools. (Learn more)

Interface Overview

Malloy Explorer is organized into three main panels:

1. Source Panel – Browse available dimensions, measures, and views from the semantic model.
2. Query Panel – Assemble and refine your query using a clean visual editor.
3. Results Panel – View results as tables or charts, and inspect the underlying Malloy and SQL.

---

2. Source Panel

The Source Panel is where exploration begins. It presents the structure of the selected Malloy model — typically organized around a central transactional or fact table — and surfaces all the fields and reusable logic available for analysis.

The panel is organized into three sections:

• Dimensions – Attributes you can group by, filter on, or sort with.
• Measures – Predefined metrics you can aggregate, filter on, or sort with.
• Views – Saved queries or perspectives authored in the model and ready to run.

Hovering over a field or view in the Source Panel will show actions you can take to add it to your query - selections will immediately appear in the Query Panel, ready to run or modify.

Dimensions

The Dimensions section lists descriptive fields — such as IDs, dates, categories, and derived attributes — from both the primary model and any joined sources.

Dimensions are grouped by their source. For example, in a query centered around order_items (the storefront fact table), you also see dimensions from its joined sources: customers, products, and regions.

Hovering over a dimension reveals contextual actions:

• Add as Group By – Segment the result by this dimension.
• Add as Filter – Apply a filter based on the field.
• Add as Sort – Sort the result set by this value.

Measures

The Measures section contains all aggregated metrics available in the current model. These are grouped by source and include calculations such as totals, averages, counts, and ratios.

Hovering over a measure provides the following actions:

• Add as Aggregate – Include the metric in the result set.
• Add as Filter – Use the measure to restrict the result set.
• Add as Sort – Sort results based on the metric value.

Views

The Views section surfaces saved queries defined in the underlying Malloy model. These views often represent curated KPIs, commonly-used explorations, or analytical building blocks. Views can be a good starting point for your queries, or may be useful to nest within your query to dive deeper.

Hovering over a view provides the following actions:

• Add to Query – Add the view to your query.
• Add as Nested Query – Add the view to your query as a nested query.

Clicking a view loads its full query into the Query Panel, where it can be run, modified, or used as a base for deeper exploration. If a query is already open, the view will be added to the query as a nested query.

---

3. Query Panel

The Query Panel is where queries come together. It provides a structured, visual representation of your query — and supports far more than basic field selection.

Top Components

• Run Button – Executes the query and displays the result.

• Clear Button – Removes the current query configuration.

• Source – Displays the source the query is running against.

• Main query

  • Add Query Element (+)
  • More Actions Menu (...) – Includes:
    • Clear Query
    • Nest Query (used to add subqueries — see below)
  • Chart Type Selector – Select the type of chart to display the results.
  • Query display – Displays the current query you've built.

Add Query Element (+)

This menu allows you to add fields and logic by operation, rather than by data source. Categories include:

• Add Group By
• Add Aggregate
• Add Filter
• Add View
• Limit
• Order By
• Add Blank Nested Query

Display Format Selector

Choose how to visualize your results. Available formats include:

• Table
• Bar Chart
• Dashboard
• Line Chart
• JSON
• List / List Detail
• Point Map
• Scatter Chart
• Segment Map

Query Configuration

Your current query is displayed using clearly labeled sections:

• Group By
• Aggregate
• Filter By
• Order By
• Limit

Each section reflects the logic of your query and is fully editable.

Hovering over a section will show another way to add elements to the query.

---

Nested Queries

Malloy's nesting feature enables rich, multidimensional analysis — and the Explorer gives you a no-code way to use it.

To add a nested query, either:

• Select Nest Query from the Query Panel More Actions menu.
• Hover over a view in the Source Panel and select Add as Nested Query.

Each nested block can have its own groupings, aggregates, filters, ordering and even additional nests.

Use nesting to:

• Break down results by subgroups
• Create expandable subtables or charts
• Explore deep, layered data relationships

Malloy compiles nested blocks into safe, aggregating subqueries in SQL. Filters can be scoped to any level.

Reference: Malloy Nested Views Documentation

---

4. Results Panel

The Results Panel displays your query output whenever you click Run. Results are rendered according to your selected chart type and include support for nested results.

Visual Output

The Results Panel renders results in the format specified by in the query above (e.g., table, chart, dashboard, etc.).

Nested queries produce expandable sub-tables or inline charts that show multidimensional breakdowns per row.

Code Transparency

Alongside visual output, the Results Panel includes: Malloy Tab – Shows the Malloy query generated by your visual configuration.

SQL Tab – Displays the fully compiled SQL sent to the database.

This transparency gives developers, analysts, and AI systems full access to inspect, reproduce, or extend any query.

---

Summary

Malloy Explorer combines semantic modeling with a powerful, no-code visual interface for data exploration. It enables you to:

• Build and run queries visually, without writing code
• Use governed metrics and reusable models for consistency
• Analyze data across multiple dimensions with nested queries
• Instantly view results as tables or charts, including expandable sub-tables
• Inspect generated Malloy and SQL code for transparency and reproducibility
• Rely on symmetric aggregates for accurate, consistent results as you drill in and out

With Malloy Explorer, everyone can ask better questions — and trust the answers.

Where to go next

• Open the Explorer from the Publisher App — pick a source in the storefront package and start building.
• Prefer plain English? Point an AI agent at the same model over MCP.
• Save what you build into a package: add the view to the model with the malloy-publish skill, or ship a no-build HTML data app.