Nebius Embedder

Generate embeddings with an active Nebius model through its OpenAI-compatible API.

Select an active embedding model from the Nebius List Models API and set its output dimensions.

import os

from agno.knowledge.embedder.nebius import NebiusEmbedder

embedder = NebiusEmbedder(
    id=os.environ["NEBIUS_EMBEDDING_MODEL"],
    dimensions=int(os.environ["NEBIUS_EMBEDDING_DIMENSIONS"]),
)
embedding = embedder.get_embedding(
    "The quick brown fox jumps over the lazy dog."
)

print(f"First values: {embedding[:5]}")
print(f"Dimensions: {len(embedding)}")

The SDK defaults to BAAI/bge-en-icl. Confirm model availability for your account instead of relying on that default, and override both id and dimensions when selecting another model.

Run the Example

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate

Export the API key and model settings

export NEBIUS_API_KEY=your_nebius_api_key_here
export NEBIUS_EMBEDDING_MODEL=your_active_embedding_model
export NEBIUS_EMBEDDING_DIMENSIONS=your_model_output_dimensions

Install dependencies

uv pip install -U agno openai

Run the example

python nebius_embedder.py

Developer Resources