Relik: Link entities and extract relations from text

Understand the project

Practical overview

Use Relik when a Python NLP workflow needs entity linking, relation extraction or a retriever-reader pipeline over text.

Developers adding entity linking or relation extraction to NLP applications.
Researchers and practitioners building retriever-reader information-extraction pipelines.

Before you choose

The library is installed with pip; the README also documents optional dependency groups for training, FAISS and serving.
The documented model-loading examples use Hugging Face model identifiers with Relik.from_pretrained.
Retriever and reader components can be used together or separately; retriever-only use returns candidates.
The README notes CPU-only availability for the PyPI FAISS package and gives separate Conda instructions for GPU FAISS.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Load the documented pretrained entity-linking model and process one text string.

Import Relik from relik.
Load the model with Relik.from_pretrained("sapienzanlp/relik-entity-linking-large").
Call the loaded model with a text string, for example relik("Michael Jordan was one of the best players in the NBA.").

The call returns a `RelikOutput` containing the input text, entity spans, and candidate entities, as illustrated in the project README.

The documented example covers one input string and the named pretrained model; it does not establish performance or results for other models or inputs.

Evidence and transparency

Source record details

Source availability: Active
Sources checked: 18 October 2026
Original website title: SapienzaNLP/relik · Official homepage title · retained in its source language