Semantic Entity Extractor (TextRank)

Discover which concepts actually dominate your text. This tool runs the TextRank algorithm in your browser — building a word co-occurrence graph, scoring it with PageRank, and surfacing the top entities, key phrases, and their connections.

AI Workflows & Technical Writing Local processing Your text stays in your browser. Nothing is uploaded.

Tool interface

Introduction

TextRank is an unsupervised keyword-extraction algorithm (the basis of many "automatic keyword" features). Running it locally lets you see your text's true conceptual focus without sending it to a cloud NLP service.

How it works

  1. Tokenize and remove stopwords.
  2. Build a co-occurrence graph over a sliding window.
  3. Run PageRank iterations to score each word.
  4. Rank entities, derive key phrases, and render the network.

Privacy

All computation is local. No upload, no account, no API key, no cloud NLP.

Limitations

  • TextRank is compute-heavy; very long documents are sampled for responsiveness.
  • Single-word entities; multi-word detection is heuristic via key phrases.

FAQ

Is this the same as TF-IDF?

No. TF-IDF needs a corpus; TextRank works on a single document using a graph, which is why it runs fully client-side.

Is my text uploaded?

No. Everything runs in your browser.

Related tools

References

  • Mihalcea & Tarau (2004) — TextRank: Bringing Order into Texts

Version and changelog

  • Version: 1.0.0
  • Last updated: August 12, 2026
  • Changelog: Initial production release.