superlinked 37.5.2

The Superlinked vector computing library

SIE: Superlinked Inference Engine is an open-source inference server for small AI models: embeddings, reranking, and extraction.
It runs on your own infrastructure, from a laptop to a production Kubernetes cluster.
The legacy superlinked Python framework is deprecated for new projects. Use sie-sdk instead.

Superlinked Python Framework (Deprecated)

Deprecated for new projects

The superlinked Python framework is deprecated for new projects.

Superlinked now focuses on SIE: Superlinked Inference Engine, an open-source inference server for small AI models. SIE runs encoders, rerankers, and entity extractors on your own infrastructure, from a laptop to a production Kubernetes cluster.

This package does not install SIE. It remains available only for existing projects that already depend on the legacy Superlinked framework.

New projects should use SIE instead:

pip install sie-sdk

SIE GitHub: https://github.com/superlinked/sie
SIE docs: https://superlinked.com/docs
Superlinked website: https://superlinked.com

Use SIE for new projects

SIE is the current Superlinked inference product for semantic search, retrieval, RAG, reranking, and information extraction workloads. It exposes three primitives:

• encode converts text or images to vectors for semantic search and RAG.
• score reranks query-document pairs for higher-precision retrieval.
• extract pulls entities and structured data from unstructured text.

Use SIE when you need to:

• Serve 85+ supported models from one API.
• Generate dense, sparse, multi-vector, and vision embeddings.
• Rerank candidate documents or items with reranking models.
• Extract entities and structured information from text.
• Run inference locally with Docker.
• Deploy inference on production Kubernetes infrastructure.
• Integrate inference into search, RAG, and document-processing systems.

SIE is not a drop-in replacement for the deprecated superlinked framework. It is the new Superlinked inference product for new applications and infrastructure.

Quickstart with SIE

SIE runs as a Docker container. Your application calls it over HTTP. Start the server, install the SDK, then call encode, score, or extract.

SIE's primary target is x86 Linux nodes with NVIDIA GPUs. The CPU image is useful for trying SIE locally.

1. Run the engine

# macOS (Apple Silicon)
docker run --platform linux/amd64 -p 8080:8080 \
  -v sie-hf-cache:/app/.cache/huggingface \
  ghcr.io/superlinked/sie-server:latest-cpu-default

# Linux (CPU)
docker run -p 8080:8080 \
  -v sie-hf-cache:/app/.cache/huggingface \
  ghcr.io/superlinked/sie-server:latest-cpu-default

# Linux (NVIDIA GPU)
docker run --gpus all -p 8080:8080 \
  -v sie-hf-cache:/app/.cache/huggingface \
  ghcr.io/superlinked/sie-server:latest-cuda12-default

The server starts on port 8080. Models are available through the API and load on first request.

Optional readiness check:

curl http://localhost:8080/readyz
# "ok"

2. Install the SDK

pip install sie-sdk           # Python
pnpm add @superlinked/sie-sdk # TypeScript

3. Generate embeddings

from sie_sdk import SIEClient
from sie_sdk.types import Item

client = SIEClient("http://localhost:8080")

result = client.encode("BAAI/bge-m3", Item(text="Hello world"))
print(result["dense"].shape)  # (1024,)

4. Rerank search results

query = Item(text="What is machine learning?")
items = [
    Item(text="Machine learning uses algorithms to learn from data."),
    Item(text="The weather is sunny today."),
]

result = client.score("BAAI/bge-reranker-v2-m3", query, items)

for entry in result["scores"]:
    print(f"Rank {entry['rank']}: score={entry['score']:.3f}")

5. Extract entities

result = client.extract(
    "urchade/gliner_multi-v2.1",
    Item(text="Tim Cook is the CEO of Apple."),
    labels=["person", "organization"],
)

for entity in result["entities"]:
    print(f"{entity['label']}: {entity['text']}")

What changed?

The superlinked package originally provided a Python framework for building vector search and recommendation applications with declarative schemas, indexes, query-time weighting, and vector database integrations.

That legacy framework is now deprecated for new projects.

Superlinked now focuses on SIE: Superlinked Inference Engine. SIE is designed for teams that need a unified inference layer for modern search, retrieval, RAG, reranking, and information extraction workloads.

Should I still use this package?

Use superlinked only if you are maintaining an existing application that already depends on the legacy framework.

Do not start new projects with this package. For new development, use SIE.

Links

• SIE GitHub: https://github.com/superlinked/sie
• SIE docs: https://superlinked.com/docs
• SIE quickstart: https://superlinked.com/docs/quickstart
• SIE HTTP API reference: https://superlinked.com/docs/reference/api
• SIE models: https://superlinked.com/models
• SIE examples: https://superlinked.com/docs/examples
• Python SDK on PyPI: https://pypi.org/project/sie-sdk/
• Superlinked website: https://superlinked.com