Skip to main content
View source

Google Cloud Storage

View as Markdown

A tool node for interacting with Google Cloud Storage buckets.

Authentication

This node uses the shared RocketRide Google Cloud authentication logic. You can authenticate by either:

  1. Uploading a Service Account JSON Key.
  2. Leaving the key blank and relying on Application Default Credentials (ADC) if your RocketRide server is hosted on Google Cloud Run, GKE, or a Compute Engine VM with attached service accounts.

Usage

When placed in a pipeline as a tool node, the LLM can:

  • list_files: List objects in the configured bucket (with optional prefix filtering).
  • download_file: Download an object from the bucket to a temporary local file on the server.

Downloaded files are written under a temp path returned as local_path. Only the most recent download is retained; starting a new download deletes the previous temp file. Remaining files are also deleted in endGlobal when the node shuts down. Objects larger than maxDownloadBytes (default 50 MiB) are rejected before download, and the downloaded size is checked again after the fetch in case the object changed.

Schema

FieldTypeDescriptionDefault
gcs.bucketNamestringBucket Name
The Google Cloud Storage bucket name.
gcs.maxDownloadBytesintegerMax Download Bytes
Reject downloads larger than this many bytes (default 50 MiB). Protects server disk from oversized objects.
52428800
gcs.prefixstringPath Prefix
Optional path prefix inside the bucket.
gcs.profilestring"default"

Dependencies

  • google-auth >=2.23.3
  • google-cloud-storage >=2.13.0