BonnardBonnard0.3.14

Getting Started

Install the Bonnard CLI, connect your data warehouse, and define your first semantic layer with cubes and views in under five minutes.

What is Bonnard?

Bonnard is a semantic layer for AI agents, applications, and dashboards. Define your metrics and dimensions once, then every consumer — Claude, Cursor, your React app, a hosted dashboard — queries the same governed data. One deploy gives you an MCP server, REST API, and SDK, all backed by the same schema.

Set up your project (with your agent)

Install the CLI

Install the Bonnard CLI globally. · view source

$npm install -g @bonnard/cli

Initialize your project

Scaffold your project. This sets up agent rules and the get‑started skill for Claude Code.

$bon init

Launch Claude Code

Start Claude Code in your project directory. It picks up the Bonnard rules and skills automatically.

$claude

Build your semantic layer

Trigger the built-in skill to have Claude walk you through connecting a data source, creating cubes, and deploying.

Prompt
help me set up my semantic layer using bonnard
Skill
/bonnard-get-started

Next steps

  • Learn about cubes — measures, dimensions, joins
  • Learn about views — curated interfaces for consumers
  • Set up MCP — connect AI agents to your semantic layer
  • Read the CLI guide — validate, deploy, query

Overview

Define your metrics once. Query governed data from any AI agent, dashboard, or application.

Part 1: Project Setup

Set up the Bonnard CLI, connect a data source, and scaffold your first semantic layer project.

On this page

What is Bonnard?Next steps