graphify

Yes, you can generate GraphRAG-ready JSON from unstructured files. Graphify processes mixed inputs like documents and papers, applying semantic extraction and community detection to export interactive HTML and JSON outputs optimized for retrieval pipelines.

GraphRAG with graphify handles inferred relationships by labeling every edge in an audit trail as EXTRACTED, INFERRED, or AMBIGUOUS, ensuring you can distinguish between explicitly found connections and guessed semantic links within your reading list.

Yes, you can process audio and video files for semantic extraction. Graphify optionally supports these formats via transcription, integrating them with documents and papers to build a complete clustered knowledge graph with community detection.

Graphify supports Neo4j export and writes Obsidian-friendly reports. It converts your raw folders into interactive HTML for browsing and GraphRAG-ready JSON, allowing you to export and explore the generated clustered knowledge graph externally.

A limitation of building a knowledge graph from a large research corpus is that processing requires chunked agent-based semantic extraction for non-code content, which may involve heavy computation before generating the final persisted graph artifact and reports.