GraphGPT turns unstructured text into an interactive knowledge graph. Paste in a movie synopsis, a dense Wikipedia passage, or a video transcript, and it extracts the entities and relationships and draws them as a visual graph you can explore.
GraphGPT is an open-source web app built around GPT-3. A carefully designed few-shot prompt guides the model to return entities and relationships in a strict JSON format, which the app then renders with a graph visualization library. Follow-up queries can update the current graph, injecting new nodes and edges or restyling existing ones, so the graph grows as you feed it more text. What started as a weekend experiment became one of the early, widely shared demos of LLM-powered knowledge extraction, and the repository has gathered over 4,400 stars on GitHub under the MIT license. The project was last updated in 2024, and you will need your own OpenAI API key to run it, since all extraction calls go to OpenAI’s models.
GraphGPT is free and open source under the MIT license. The only cost is your own OpenAI API usage, which is billed by OpenAI according to the model you configure.
Students, researchers, and developers who want a quick visual way to understand relationships hidden in text, or a small reference implementation of LLM-based knowledge graph extraction.