Embedding Providers¶

Vector Graph RAG supports explicit embedding provider selection. New applications should set both embedding_provider and embedding_model instead of relying on model-name inference.

Supported Providers¶

Provider	Extra	Typical use
`openai`	built in	OpenAI or OpenAI-compatible embedding endpoints.
`huggingface`	`hf`	Local HuggingFace transformers models.
`google` / `gemini`	`google`	Google Gemini embeddings.
`voyage`	`voyage`	Voyage AI embeddings.
`jina`	`jina`	Jina AI embeddings.
`mistral`	`mistral`	Mistral embeddings.
`ollama`	`ollama`	Local Ollama embedding server.
`local`	`local`	Sentence Transformers style local embeddings.
`onnx`	`onnx`	ONNX Runtime embedding models.

Install only the provider extras you need:

uv add "vector-graph-rag[jina]"
uv add "vector-graph-rag[ollama]"
uv add "vector-graph-rag[all]"

Configure A Provider¶

from vector_graph_rag import VectorGraphRAG

rag = VectorGraphRAG(
    embedding_provider="openai",
    embedding_model="text-embedding-3-small",
)

For providers that need a separate key or endpoint, use embedding_api_key and embedding_base_url:

rag = VectorGraphRAG(
    embedding_provider="jina",
    embedding_model="jina-embeddings-v4",
    embedding_api_key="jina_...",
)

rag = VectorGraphRAG(
    embedding_provider="ollama",
    embedding_model="nomic-embed-text",
    embedding_base_url="http://localhost:11434",
)

Dimension Consistency¶

All entities, relations, and passages in one graph must use the same embedding model and dimension. If you change embedding providers or embedding dimensions, rebuild that graph or use a new collection_prefix.

rag = VectorGraphRAG(
    collection_prefix="finance_openai_small",
    embedding_provider="openai",
    embedding_model="text-embedding-3-small",
)

Legacy Provider Inference¶

If embedding_provider is omitted, Vector Graph RAG keeps a legacy compatibility path that infers the provider from embedding_model. That path is deprecated and planned for removal in v1.0.0. Prefer explicit configuration:

rag = VectorGraphRAG(
    embedding_provider="openai",
    embedding_model="text-embedding-3-small",
)

See the Python API Reference for constructor parameters and provider examples.