bookstrap-ingest

To ingest research PDFs and Markdown files into a knowledge graph, the Skill performs semantic chunking, generates embeddings, extracts entities, and builds relationship structures from your provided sources. It processes files, directories, and URLs to populate a centralized research corpus.

You need semantic chunking and entity extraction when preparing diverse research materials for long-form writing. This process converts unstructured PDF, HTML, and text sources into a searchable database by identifying characters, locations, and events, and mapping their relationships.

Yes, you can build a research timeline from unstructured HTML and text documents. The Skill extracts entities and sequences events with their associated dates during ingestion, constructing a chronological timeline to support storytelling and historical validation.

Yes, you need a SurrealDB instance configured alongside an embedding provider. The database stores the extracted chunks, generated embeddings, entities, and relationship structures required to manage your research corpus effectively.

The best way to convert multiple research sources into a structured database is through automated ingestion that handles semantic chunking and relationship construction. This approach transforms scattered files and URLs into a centralized, searchable knowledge base ready for analysis.

Ingesting research sources supports PDF, Markdown, HTML, and plain text formats, but requires configured embedding providers and a SurrealDB instance. Sources outside these specified formats or without proper database configuration cannot be processed into the structured knowledge graph.