video1.0<iframe src="https://www.loom.com/embed/c1822582396d4704a1a54901f8995bf0" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/c1822582396d4704a1a54901f8995bf0-4ef328429201fefb.gif248.531This Loom demonstrates an event-driven workflow system that builds a knowledge graph from incoming Telegram links and prevents duplicate ingestions. When a link is sent, it is first saved to the database, and if the message URL already exists it reports that it is already in the knowledge graph. It then triggers subsequent workflows to extract content and run a GraphBuilder, producing a report with 52 nodes and 38 edges for the example article. The speaker notes the project is about 70% complete and is planning to build an agentic system on top of the knowledge graph.