Thanks for sharing the traceback — that error usually means you’re running a version of

prefect-dbt

that was built against Pydantic v1, but your environment has Pydantic v2 (which Prefect 3.x uses). Pydantic v2 removed

__modify_schema__

, so older libraries that still reference it will raise this error during import. A couple of quick checks and fixes: 1) Check your versions Run this to see exactly what’s installed:

python -c "import sys, prefect, prefect_dbt, pydantic; print('python', sys.version); print('prefect', prefect.__version__); print('prefect_dbt', getattr(prefect_dbt, '__version__', 'unknown')); print('pydantic', pydantic.__version__)"

2) If you’re on Prefect 3.x (likely on Python 3.12) Upgrade

prefect-dbt

to the latest release that supports Pydantic v2: - pip:

pip install -U prefect prefect-dbt

- conda-forge:

conda install -c conda-forge prefect prefect-dbt

3) If you must stay on Prefect 2.x Prefect 2.x relies on Pydantic v1. In that case, either: - Pin Pydantic to v1:

pip install "pydantic<2"

- Or upgrade to Prefect 3.x + latest

prefect-dbt

. Why this happens: importing

prefect_dbt

triggers its

__init__

which imports the Cloud credentials module. Older

prefect-dbt

builds define schema tweaks the Pydantic v1 way, which fails under Pydantic v2. If you share the output of the version check above, I can give you exact pin recommendations for your setup. Reference: prefect-dbt integration guide