entity-topic-cluster is a browser-based tool that groups topics using entity recognition and machine learning. It works fully on your device. You don’t need to install anything or send data online. The tool finds important names and ideas in your text, links them to Wikidata, and shows them as a knowledge graph. This helps you understand how topics relate and find connections clearly.
This tool works on most Windows computers. It runs inside your web browser. You only need:
- A computer running Windows 10 or later
- A modern web browser (such as Chrome, Edge, or Firefox)
- An internet connection to download the tool (after that, it works offline)
- At least 4 GB of RAM for smooth performance with large files
No extra software or extensions are required. It does not depend on remote servers. All processing happens on your device.
Follow these steps to start using entity-topic-cluster on Windows:
- Open your web browser (Chrome is recommended).
- Go to the official repository page by clicking the button below.
- On the GitHub page, find the Releases section or the main page, as the tool runs in the browser.
- No installation is needed. You simply open the web app from the repository or local files.
- Prepare your text data. This can be a collection of documents, articles, or any text file.
- Open the entity-topic-cluster web app in your browser using the link on the repository.
- Upload your text or paste it into the provided input area.
- The tool will automatically find named entities (like people, places, and organizations).
- It will link these entities to Wikidata entries for more details.
- The clustered topics appear as an interactive graph on the screen.
- You can click on nodes in the graph to explore relationships and understand connections better.
- Use the controls to zoom, pan, or focus on specific parts of the graph.
- 100% client-side: Your data stays on your device.
- Entity extraction: Finds names, places, dates, and concepts.
- Wikidata linking: Connects entities to a large knowledge base.
- Topic clustering: Groups related themes automatically.
- Knowledge graph visualization: Shows data as an interactive map.
- No installation or signup required.
- Works on most Windows browsers.
- Supports large text inputs with efficient processing.
- Open source with transparent code.
If you want to run the application offline from your computer:
- Download the repository or the latest release ZIP file.
- Extract the ZIP file to a folder on your Windows PC.
- Open the extracted folder and find
index.html. - Double-click
index.htmlto open the app in your default browser. - Use the app as described above.
This method keeps the app on your device with no internet required.
- If the app does not open, make sure you use a modern browser like Chrome or Edge.
- Large files might take some time to process. Give it a moment.
- If the graph looks empty, check your input text contains enough data and entities.
- If the page shows errors, try refreshing or restarting your browser.
- For connection errors when downloading, check your internet connection.
entity-topic-cluster uses machine learning models that run inside the browser. This design protects your privacy. Since everything happens on your computer, no data is uploaded or stored elsewhere.
The tool supports natural language processing tasks like named entity recognition and topic clustering. It uses transformers-js for the models in JavaScript, making it fast and browser-compatible.
The visual graph helps you see how concepts connect and overlap in articles or collections of documents. It can assist with SEO planning or research by organizing complex information simply.
- Repository Name: entity-topic-cluster
- Description: 100% client-side entity-based topic clustering tool. ML models run in your browser — extract entities, resolve to Wikidata, visualize as knowledge graph.
- Topics: cloudflare-pages, entity-resolution, knowledge-graph, machine-learning, named-entity-recognition, nlp, seo, topic-clustering, transformers-js, wikidata
To get started, visit this page to download or run the tool in your browser: