/bookstrap-ingest - Load Research Corpus

Load initial research materials into the Bookstrap database for use during writing.

Purpose

Ingest source documents into the database to build the research corpus. This command processes multiple file types (PDF, markdown, HTML, plain text), generates embeddings for semantic search, extracts entities (characters, locations, events, concepts), builds relationships between entities, and constructs the knowledge timeline.

Input Arguments

Accept one or more of the following:

• File paths: Individual files to ingest (e.g., ./research/soe-training.pdf)
• Directories: Recursively process all files in a directory (e.g., ./research/)
• URLs: Web pages to fetch and ingest (e.g., https://example.com/article.html)

Multiple sources can be provided in a single invocation:

/bookstrap-ingest ./research/documents/ https://example.com/article.html ./notes.md