dvc-gs

What it is and what it does

dvc-gs is a plugin that extends DVC (Data Version Control) to work with Google Cloud Storage as a remote data backend. Once installed, it allows you to store and retrieve versioned datasets and models directly in GCS buckets using DVC's standard remote commands, without needing to manage separate cloud storage APIs.

The plugin sits between DVC and gcsfs (the Google Cloud Storage filesystem library) and handles the protocol translation so that gs:// URLs work seamlessly in your DVC pipelines and remote configuration. It's part of the DVC ecosystem of storage plugins and is actively maintained by Iterative.

Use it for:

• Store DVC-tracked datasets in GCS for team collaboration and CI/CD pipelines.
• Use GCS as a remote cache to avoid re-downloading large model artifacts across machines.
• Integrate DVC workflows with Google Cloud Vertex AI or other GCP-native ML services.
• Version control and track ML experiment outputs stored in GCS buckets.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Enables DVC to read from and write to Google Cloud Storage by providing a gs:// remote backend plugin.

Yes, if you use DVC and need GCS as a remote backend. Low install friction, active maintenance, no known vulnerabilities, and permissive licensing make it a straightforward addition to a DVC-based workflow. Install only if you have a GCS bucket and DVC already in use.

Install

dvc-gs on PyPI

pip install dvc-gs

uv add dvc-gs

poetry add dvc-gs

Installing dvc-gs

Before you install

Low friction install with two runtime dependencies (dvc and gcsfs). Actively maintained with recent commits and Python 3.9–3.14 support.

License in practice

Apache-2.0 permissive license allows commercial and private use with minimal restrictions.

Quickstart

pip install dvc-gs

After installation, configure a DVC remote:
dvc remote add mygs gs://my-bucket/path
dvc remote default mygs

Requires dvc to be installed and a Google Cloud project with credentials configured (via GOOGLE_APPLICATION_CREDENTIALS or gcloud auth).

Verify before relying

• Whether the plugin handles authentication methods beyond service account keys (e.g., Application Default Credentials, workload identity).
• Performance characteristics and limits for large-scale data transfers to GCS.

Package facts

Evidence: dvc_gs-3.1.0-py3-none-any.whl

Keywords: dvc, gs

Tags

More Software Development packages