---
id: keybert
version: "0.9.0"
license: MIT License Copyright (c) 2020, Maarten P. Grootendorst Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the… (full text in the JSON record)
license_treatment: permissive
maintenance: active
---

# keybert — KeyBERT performs keyword extraction with state-of-the-art transformer models.

License: permissive · Maintenance: active · Downloads: 720.9K/mo

## What it is and what it does

KeyBERT is a lightweight keyword extraction library that uses pre-trained BERT embeddings to identify the words and phrases in a document that best represent its semantic content. It works by computing a document-level embedding, then embedding candidate n-grams, and ranking them by cosine similarity to the document. The library requires four runtime dependencies: numpy and scikit-learn for numerical operations, sentence-transformers to load and run embedding models, and rich for output formatting.

The package is designed for minimal setup. It supports multiple extraction strategies, configurable n-gram ranges, and pluggable embedding backends. It is actively maintained, supports Python 3.8, 3.9, 3.10, 3.11, and 3.12, and has no known security vulnerabilities.

Use it for:

- Extract single-word keywords from research papers or articles to build document summaries or search indices.
- Generate multi-word keyphrases from product descriptions or customer feedback to identify key topics.
- Diversify keyword results using maximal marginal relevance to avoid redundant or near-duplicate extractions.
- Build a document tagger by extracting keywords and using them to label or categorize incoming text.
- Integrate semantic keyword extraction into a content recommendation system to match documents by extracted terms.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

KeyBERT extracts keywords and keyphrases from documents by computing BERT embeddings and finding n-grams most similar to the document as a whole using cosine similarity.

Yes. KeyBERT is worth installing for semantic keyword extraction tasks. It has low install friction, active maintenance, permissive licensing, no known vulnerabilities, and a straightforward API. The main consideration is that sentence-transformers downloads a transformer model on first use; verify that the model can be pre-cached and that extraction latency meets your needs.

## Install

pip install keybert
uv add keybert
poetry add keybert

## Installing keybert

Before you install: Low install friction: pure Python wheel with four runtime dependencies (numpy, scikit-learn, sentence-transformers, rich). Actively maintained with recent commits and no known vulnerabilities.

License in practice: MIT License permits commercial and private use, modification, and redistribution with minimal restrictions—suitable for most projects.

Quickstart:

pip install keybert

from keybert import KeyBERT

kw_model = KeyBERT()
keywords = kw_model.extract_keywords(doc)

sentence-transformers downloads a transformer model on first use; requires internet access or pre-cached model.

Verify before relying:

- Whether the package works offline after initial model download, or requires network access per extraction call.
- Memory footprint and latency characteristics for large documents or batch processing.
- Whether optional extras (flair, gensim, spacy, use) are required for production use or only for advanced embedding backends.
- Performance characteristics and typical extraction latency on standard document sizes.

## Package facts

- License: MIT License Copyright (c) 2020, Maarten P. Grootendorst Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the… (full text in the JSON record) (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 720.9K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

keyword extraction with BERT, automatic keyphrase extraction, document keyword mining, semantic keyword extraction, transformer-based keyword extraction, embedding-based keyword extraction, extract keywords from text, NLP keyword identification, nlp, embeddings, information-extraction

[View on SkillFed](https://skillfed.io/packages/keybert) · [View on PyPI](https://pypi.org/project/keybert/)