Skip to content

Try with a sample project

Generate real manifest.json and run_results.json from a dbt run on fictional jaffle data—no warehouse account required for the default path. Then point any dbt-tools command at ./target in the project directory.

For what each artifact file means and when dbt writes them, see dbt artifacts & target/.

This dbt Labs sample runs locally on DuckDB. Use its README for the latest Python and dbt version notes.

Prerequisites: Node.js 20+ (dbt-tools); Python 3.x and pip install -r requirements.txt per the jaffle repo; dbt 1.10+ for supported artifact schemas (version compatibility).

bash
git clone https://github.com/dbt-labs/jaffle_shop_duckdb.git
cd jaffle_shop_duckdb
python3 -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate
pip install -r requirements.txt
dbt build

From the same directory (artifacts at ./target):

bash
npx @dbt-tools/cli status --dbt-target ./target --json
npx @dbt-tools/cli discover --dbt-target ./target "orders" --limit 5 --json
npx @dbt-tools/web --dbt-target ./target

Discover examples such as orders or customers match jaffle model names.

Your own dbt project

If you already have a dbt project, run any command that produces artifacts, then point dbt-tools at target/:

bash
cd my-dbt-project
dbt build   # or dbt run
npx @dbt-tools/cli status --dbt-target ./target --json

Optional — jaffle-shop

The classic jaffle-shop tutorial project is another public sample. It requires a configured warehouse profile (Snowflake, BigQuery, and so on)—not the zero-setup DuckDB path. Use it when you already have adapter credentials and want the original tutorial layout.

Use with the quickstart

Every command in the 5-minute quickstart works once ./target contains manifest.json and run_results.json from dbt build (or dbt run) in the sample repo directory.

Public screenshots and demos

For screenshots, blog posts, or shared coding-agent sessions:

  • Prefer artifacts from public sample projects (jaffle) or sanitized non-production targets.
  • Never use production artifacts—they may expose schema paths, SQL, warehouse identifiers, or environment metadata.
  • Jaffle data is fictional, but artifacts are still real dbt JSON and may contain paths or SQL in error messages. See Data boundaries.

Released under the repository license terms.