Text2graph is an online tool for graph extraction from free text. It uses state-of-the-art LLMs from OpenAI to detect entities and their relationships. Created graph can be easily modified by renaming or merging nodes and relationships. These actions are then used to teach the algorithm to make more precise results in subsequent runs. Graphs can be exported as Cypher or Gremlin statements, CSV files, or PNG images.
Follow these steps to get the most out of the Text2graph tool
Import a file (Word, PDF) or paste the first paragraph of your text. You can import up to 10.000 characters per paragraph. Splitting your text into paragraphs is important for the continuous learning mechanism and also for better navigation in the final graph to find which entities are mentioned in which parts of the text.
Take a look at the generated graph, find similar entities, and merge them. Similarly, you can merge duplicate relationships. This process is very important because merged or changed names are stored as synonyms and used in the next paragraphs to get better results from AI.
Continue with importing the next paragraphs, merging and renaming nodes and relationships after every import. As your history grows you can easily find which nodes are mentioned in multiple paragraphs by clicking on a node. Similarly, you can see all nodes and relationships imported from a single paragraph by clicking on the paragraph in the history panel.
When your project is finished you can export the graph as Cypher or Gremlin statements to quickly import the graph into your graph DB. Or you can export it as CSV files, or PNG image. To import the data into a graph DB you can use Graphlytic, our graph analytics and automation platform.
Demtec, s.r.o.
Hraničná 18
821 05 Bratislava
Slovakia
ID: 47807890
VAT ID: SK2024104434
EMAIL: info@graphlytic.biz
TEL: +421 944 289 809
