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/.
Recommended — jaffle_shop_duckdb
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).
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 buildFrom the same directory (artifacts at ./target):
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 ./targetDiscover 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/:
cd my-dbt-project
dbt build # or dbt run
npx @dbt-tools/cli status --dbt-target ./target --jsonOptional — 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.