Alembic

Vertex AI Workspace for Terpedia

Set up a Workbench, ask questions, and generate reports with Gemini on Google Cloud.

Setup & launch a Workspace

Use a Vertex AI Workbench (Jupyter) instance in your GCP project, then run the Alembic client inside it.

  1. Create a Workbench instance (if you don’t have one): Open Vertex AI Workbench → Instances and click Create new, or run: 
    gcloud workbench instances create terpedia-workbench \
  --project=YOUR_PROJECT_ID \
  --location=us-central1-b \
  --vm-image-project=cloud-notebooks-managed \
  --vm-image-family=workbench-instances \
  --machine-type=n1-standard-4
  2. Open JupyterLab from the instance row (Open JupyterLab) in the console.
  3. In a JupyterLab terminal, clone and install Alembic: 
    git clone https://github.com/Terpedia/alembic.git
cd alembic
pip install -r requirements.txt
pip install -e .
  4. Set GOOGLE_CLOUD_PROJECT in the notebook (or in a .env file in alembic/), then use the client or CLI (see below).

View repo Open Workbench console

Ask questions & write reports

Enter a question or report topic. You can run the generated command in your Workbench terminal, or (if configured) use an API endpoint.