Skip to content

Getting Started


Recce requires that your dbt project has two environments to compare. For example, one for production and another for development.

Prepare two targets with separate schemas in your dbt profile. Your profiles.yml might look something like this:

  target: dev
      type: duckdb
      path: jaffle_shop.duckdb
      schema: dev
      type: duckdb
      path: jaffle_shop.duckdb
      schema: main

Install Recce

Install Recce using pip:

pip install -U recce

Use Recce in your dbt project

The following instructions give an overview of the process of using Recce in your dbt project. For a hands-on tutorial, please check the Jaffle Shop Tutorial.

Navigate to your dbt project.

cd your-dbt-project/

Prepare dbt artifacts

Recce expects two sets of dbt artifacts to be present:

  • target-base/ - dbt artifacts for to be used as the base for the comparison e.g. production
  • target/ - dbt artifacts for your development branch

Generate artifacts for the base environment

Checkout the main branch of your project and generate the required artifacts into target-base. You can skip dbt build if this environment already exists.

git checkout main

dbt run --target prod
dbt docs generate --target prod --target-path target-base/

Generate artifacts for the target environment

git checkout feature/my-awesome-feature

dbt run
dbt docs generate

Start the Recce server

Start the Recce server with the follow command:

recce server

Recce use dbt artifacts, which is generated when every invocation. You can find these files in the target/ folder.

artifacts dbt command
manifest.json dbt docs generate, dbt run, ..
catalog.json dbt docs generate


The regeneration of the catalog.json file is not required after every dbt run. it is only required to regenerate this file when models or columns are added or updated.