Build Bounded Document Ingestion Pipelines
Edit on GitHubScan selected document formats, collect ingestion summaries, and feed structured chunks into downstream indexing or analysis systems.
Read-OfficeDocument -As Ingest combines folder discovery and document reading into one bounded operation. It reports files scanned, files parsed, chunks produced, and errors instead of forcing every caller to rebuild the same orchestration.
Select the corpus
$result = Read-OfficeDocument -As Ingest -Path '.\Knowledge' -Recurse `
-Extension docx,pdf,md,html,json,yaml `
-MaximumFiles 1000 -MaximumTotalBytes 2GB
Omit -Recurse for a single controlled directory. Extension filters keep unrelated files outside the parser path. Reader options and content bounds should match the same policy used by search and chunk extraction.
The ingest-folder recipe ingests supported documents from a simple .\Documents folder.
Keep indexing, embeddings, databases, or network publication outside the Reader command. Reader owns deterministic document extraction; the downstream system owns storage and retrieval policy.