WHY THIS
EXISTS.
Paperless-ngx is excellent at storing and OCRing documents, but meaning-based retrieval, receipt intelligence and accounting workflows still require manual work or sending private paperwork to cloud AI services.
OSS / LOCAL-FIRST · AI & models
// Your paperwork grows a cortex
PaperCortex
A private intelligence layer for Paperless-ngx: semantic search, classification, receipt extraction and accounting export powered entirely by local models.
LAST PUSH…syncing with GitHub
LANGUAGE…primary repository language
RELEASENO FORMAL RELEASElatest public GitHub release
LICENSENOT DECLAREDrepository metadata
WHY THIS
THE USEFUL
LOOP.
PaperCortex keeps the intelligence layer beside the archive. Local Ollama models classify documents, generate embeddings and extract structured data, while MCP exposes the results to controlled agent workflows.
03 // WHAT_SHIPPEDWHAT'S
WHAT'S
INSIDE.
SEARCH BY MEANING
Local embeddings find related concepts, synonyms and translations instead of relying on exact OCR keywords.
RECEIPTS TO ACCOUNTING
Vendor, tax, amounts and line items become structured data with matching and DATEV or CSV export.
DOCUMENTS AS TOOLS
An MCP server makes the archive queryable by compatible agents without moving the source documents to the cloud.