November 19th, 2025 - YouTube Live
This is supporting material for the Neo4j Live: Build Smarter Graph Visualizations with AI Agents videocast.
Sebastian shows two major ways how to use AI agents and LLM tools to build graph visualizations.
For this he is mostly using two different visualization tools:
- yFiles Graphs for Streamlit - a free Python widget for the streamlit platform to render graphs from various datasources. (GitHub)
- yFiles for HTML - a commercial JavaScript library to render graphs in the browser.
- Online Playground - no sign-up required. There is a Neo4j-specific Example included among others. Here is an almost minimal one specific for this task.
- Demos - several hundred graph visualization demos, including the Neo4j-specific ones.
- GitHub Demo Sources. Contains the sources for the demos in the package
- Documentation - comprehensive documentation and examples for the various yFiles family members.
- More yFiles/Neo4j specific material
He also briefly uses the free DataExplorer for Neo4j tool from yWorks to briefly explore Neo4j data.
The tiny Streamlit app that he develops is documented here
The one-shot prompt conversation he did with OpenAI's ChatGPT-5.1 (with web search enabled!) is here. The refinement-run of the conversation is here.
The advanced application he develops is here. For this he uses the neo4j-mcp Server and the yFiles MCP Server with yFiles for HTML for the visualization. Evaluate it for free from here. The IDE he uses is the free VSCode with the free Cline Extension as the coding agent. As models he uses OpenAI's GPT series (5.1, 4.1 mini/nano), Anthropic's Claude Code and Haiku 4.5, the (currently free) grok-code-fast-1.
See the coding agent in action in these YouTube Shorts from the yWorks Channel:
Get in contact with yWorks - the maker of yFiles - here