What’s in this Quick Start?
This Quick Start guides you through the end-to-end Credible workflow, focusing on a familiar and powerful pattern: a data modeler defines trusted business logic in VS Code using the Credible AI agent, and a data analyst explores the governed data model using Malloy’s no-code Explorer query builder. This mirrors the proven modeler → analyst workflow popularized by tools like DBT and Looker — now reimagined for the AI era. You will learn how to:- Build a semantic model using the Credible AI agent in VS Code
- Create a view to validate your model
- Publish your work to the Credible service for broader use
- Explore your semantic model using the no-code Explorer to answer business questions
Step 0: Get Set Up
Before starting, make sure:- A Credible admin set up your organization. (self-service demo coming soon)
- You have a basic understanding of Malloy, the semantic modeling language Credible is built on. View Malloy Docs →
Install the CLI and Log In
Open a terminal and install the Credible CLI globally:<yourOrgName> with the organization name provided by your admin):
Install the VS Code Extension
- In VS Code, go to the Extensions view (
Cmd+Shift+X) - Search for
Credibleand install the extension
Credible Service Panel is where you can view your organization’s projects, connections, and packages. For example:
quick-start project, you can see you have access to a “bq_demo” connection. The connection is managed by the Credible service and is not stored locally, but the Credible VS Code extension will discover it in your project and allow you to use it in your local models—thus eliminating the need to store database credentials locally.
Step 1: Build a Semantic Model
At the core of Credible is the semantic model—a governed, versioned interface that defines how your data should be understood and used. Think of it as a an API for your data: it captures not just structure, but business meaning. You’ll use the Credible AI agent in VS Code to generate your first model using existing catalog metadata, query logs, and the structure of your data warehouse.Open the Credible Chat Panel
Open the Credible Chat Panel. You will initially see it on the left in Activity Bar. To open the Credible AI agent chat, click on the Credible icon. To view the File Explorer, the Credible Service Panel and the Chat (along with your file), we recommend moving the chat over to the right panel by dragging and dropping the Credible Chat into a panel on the right, as shown in the screencast below:Generate Your Semantic Model
Open the Credible chat and ask the Credible AI agent to help you build a model on the ecommerce dataset. For example: Build a semantic model of our ecommerce data to analyze sales by brand or product and user spend by age bucket.Review and Adjust Your Model
After the model is generated, review and approve the edit to move the model into your malloy file. The ecommerce dataset has four tables:order_items, users, products, and inventory_items. You should see Malloy sources for each table with dimensions, measures, and join relationships connecting them logically.
Step 2: Create a View
Creating views is a powerful way to validate your semantic model and define reusable analyses. We’ll use the Credible AI agent’s autocomplete feature to create a simple view that calculates sales by brand.Step-by-Step Instructions:
- Find the
order_itemssource in yourecommerce.malloy - Below the dimensions and measures, start typing
view: sales_by_brandand pause (be sure to include a space after brand to trigger the copilot) - Wait for the copilot suggestion - it should suggest a complete view definition with a measure for total sales by brand. Adjust as desired & resolve any errors
- Add a
# bar_charttag above the view to make it display as a bar chart - Click “Run” above the view execute the query. Results will display on the right
Create a Nested View
You can also create nested views to drill down into your data:Troubleshooting
If autocomplete isn’t working: add a space/tab after typing, wait 2-3 seconds, and ensure you start with valid Malloy syntax likeview: .
Step 3: Publish to the Credible Service
Now that you have a semantic model, you can publish it so it can be served to consumption experiences—AI agents, applications, and dashboards. The unit of publishing in Malloy and Credible is a package: a directory containing apublisher.json file along with .malloy models (and optionally .malloynb notebooks or embedded data files). For best practices on organizing projects and packages, see Projects & Packages Best Practices.
We’ll use the CLI to publish this package:
-
Go back to the terminal (press
Ctrl + `on Windows/Linux orCmd + `on Mac) -
Navigate to your package directory:
-
Publish the model:
The
--set-latestflag pins this version as the default “latest” version for consumers. Omit this flag to publish without pinning. -
Confirm the deployment: Refresh the “Credible Service” panel in VS Code to verify that version
0.0.1of your model now appears
Voila!
Your semantic model is now:- Governed and version controlled
- Discoverable and queryable via APIs and notebooks
- Ready for use by analysts, embedded apps, dashboards, or AI agents
Step 4: Explore your Semantic Model
Now that you’ve published your semantic model to the Credible platform, it’s time to see how data consumers can analyze data in the no-code Explorer UI (built on Malloy’s open source Publisher & Explorer functionality). This is where your work as a data modeler becomes a launchpad for analysts, product managers, and other data consumers to ask meaningful questions — without writing a single line of SQL.Open the Explorer and Run a Saved View
-
Access your model: Navigate to
https://<your-org>.admin.credibledata.com/quick-start, log in, and open yourecommercepackage. Go to Models →ecommerce.malloyand select theorder_itemssource tab. -
Run an analysis: In the Views section, click on
sales_by_brand, then click “Run” to execute the query.
Create a Nested Query
Now let’s build a query from scratch using clicks: Clear the query panel, addbrand as Group By and total_sales as Aggregate, then click + to add a nested query with user_age_group and total_sales.
✅ You’ve just created a nested query that breaks down sales by brand and age group.
🎉 What You’ve Accomplished
✅ Defined business logic in a semantic model (with AI assistance)✅ Validated the model and saved a view
✅ Published the package organization-wide
✅ Explored data with nested analysis through a no-code interface (no SQL required) You’ve experienced the power of Credible: one semantic model that works everywhere — from modeling to exploring, enabling developers and analysts to work together on the same trusted foundation.