Google Vertex AI embedder

Embeds through the Vertex AI text-embedding SDK using ambient GCP auth. Selected via EMBEDDING_PROVIDER=vertexai.

Enable it

EMBEDDING_PROVIDER=vertexai
VERTEX_PROJECT=your-gcp-project
VERTEX_REGION=us-central1
VERTEX_EMBEDDING_MODEL=text-embedding-004
EMBEDDING_DIMS=768

pip install google-cloud-aiplatform (imported lazily only when this provider is selected).

Configuration

Variable	Default	Description
EMBEDDING_PROVIDER	openai	Set to vertexai.
VERTEX_PROJECT	none	GCP project passed to vertexai.init.
VERTEX_REGION	us-central1	GCP location passed to vertexai.init.
VERTEX_EMBEDDING_MODEL	text-embedding-004	Vertex embedding model.
EMBEDDING_DIMS	768	Vector dimension; must match the embedding model.

Notes

• Credentials come from ambient GCP auth (application default credentials or workload identity), so there is no API key.
• The vertexai SDK is heavy and needs GCP auth, so it is imported and initialized lazily only when the provider is actually selected.
• If google-cloud-aiplatform is not installed, selecting this provider raises an ImportError with the install hint.