Agent-first, human-friendly business intelligence

Docs · Data Talks

---

Rill is the fastest BI tool for humans and agents, powered by OLAP engines like ClickHouse and DuckDB.

Get Started

curl https://rill.sh | sh # install
rill start my-project # create a project and open the UI

Scaffold a project with agent context

Use rill init to scaffold a project interactively:

➜ rill init
? Project name my-rill-project
? OLAP engine duckdb
? Agent instructions claude
Created a new Rill project at ~/my-rill-project
Added Claude instructions in .claude and .mcp.json
Success! Run the following command to start the project:
rill start my-rill-project

Why Rill?

• Build with agents — BI-as-code (YAML + SQL) means coding agents like Claude Code and Cursor can author projects, dashboards, and security policies end-to-end
• Semantic layer — Single source of truth for dimensions, measures, and time grains — defined in YAML, generating SQL at query time against your OLAP engine
• Explore with agents — Conversational BI lets business users query metrics in natural language; the MCP server connects AI agents directly to your semantic layer
• Real-time performance — Sub-second queries at any scale; ClickHouse for billions of rows, DuckDB for smaller datasets and fast iteration
• Embeddable — Dashboards, APIs, and agent interfaces you can ship in your product

Capabilities

Rill Developer (local)

• Connectors — S3, GCS, databases, and 20+ sources
• OLAP Engines — Managed ClickHouse or DuckDB included, or connect an external engine (ClickHouse Cloud, Druid, Pinot, MotherDuck)
• SQL Models — Transform raw data with SQL, join models together
• Data Profiling — Instant column stats and distributions
• Incremental Ingestion — Load only new data on each run to keep large datasets current without full refreshes
• Semantic Layer — Dimensions, measures, and time grains in YAML
• Row Access Policies — Per-user, per-group data access control
• Local Dashboards — Preview and explore dashboards locally

Rill Cloud

• Deploy — Git-backed, versioned deployments — push with rill deploy or connect a repo for automatic CI/CD
• Explore & Canvas Dashboards — Interactive dashboards, embeddable in your product
• Conversational BI — Ask questions in natural language
• MCP Server — Connect Claude, ChatGPT, or any AI agent to your metrics
• Custom APIs & Embedding — Expose metrics via REST or embed dashboards
• Alerts & Reports — Threshold alerting, code-defined or UI-defined

How It Works

Define everything in code — models, metrics, dashboards — and Rill handles the rest.

1. Connect data — models/events.yaml

type: modelconnector: duckdbmaterialize: truesql: | select * from read_parquet('gs://rilldata-public/auction_data.parquet')

2. Define metrics — metrics/events_metrics.yaml

version: 1type: metrics_viewmodel: eventstimeseries: timestampdimensions:
- name: countrycolumn: country
- name: devicecolumn: device_typemeasures:
- name: total_eventsexpression: count(*)
- name: revenueexpression: sum(price * quantity)description: Total revenue

3. Create a dashboard — dashboards/events_explore.yaml

type: exploredisplay_name: "Events Dashboard"metrics_view: events_metricsdimensions: "*"measures: "*"

4. Deploy

rill deploy # push to Rill Cloud

Your metrics view is immediately queryable on Rill Cloud — add YAML files to configure dashboards, alerts, and custom APIs.

Learn More

Getting Started with Rill Developer • Exploring Data with Rill • Data Talks on the Rocks • Agentic Analytics with Claude Code and Rill

Examples

Example	Description	Links
Programmatic Ads	Bidstream data for pricing and campaign performance	GitHub · Demo
Cost Monitoring	Cloud infra merged with customer data	GitHub · Demo
GitHub Analytics	Contributor activity and commit patterns	GitHub · Demo

Or explore a live embedded dashboard.

Community

Contributing

We welcome contributions! See our Contributing Guide to get started.