seed

Ingest external content into a structured knowledge management inbox by converting URLs, files, and pasted text into traceable items. The process writes YAML frontmatter with source, author, date, and reason to ensure full provenance tracking for research curation.

Deduplicate ingested research data by hashing the raw content during the seeding process. This hashing mechanism identifies identical inputs across varied sources, preventing duplicate inbox entries while preserving the original source metadata for audits.

Batch seeding multiple URLs and files into the ingestion pipeline is supported. You can process varied inputs across different sources simultaneously, automatically generating structured inbox items with provenance data for each external input.

Provenance metadata for clipboard inputs captures the source, author, and date. The ingestion workflow writes this data into YAML frontmatter within the structured inbox item, ensuring that even pasted clipboard text remains fully traceable for knowledge curation.

Configure the inbox storage path for your knowledge curation pipeline by setting the target directory in the seed workflow. The system stores deduplicated items with their YAML frontmatter metadata under this configured inbox path.

Supported input types for traceable knowledge ingestion include URLs, local files, pasted text, and clipboard data. The system converts these varied raw content sources into structured inbox items with provenance metadata.