complydoc

Pipeline

Files are read locally by per-format loaders, then passed to three analysis components that run independently. Each run writes a JSON report and an HTML report.

complydoc pipeline architecture A folder of documents passes through discovery and a format-specific ingest layer into three independent analysis components — cost, difficulty and sensitive data — which are driven by YAML config and emit a diffable JSON report and a self-contained HTML report. The whole pipeline sits inside a network guard boundary, so no document content can leave the machine. NETWORK GUARD ANALYSIS · INDEPENDENT THRESHOLDS INPUT documents/ pdf · img · docx · xlsx WALK Discovery skips unopenable LOAD Ingest per-format loaders 01 Cost tokens, pages, price 02 Difficulty 18 measured signals 03 Sensitive masked by default YAML Config prices · weights · patterns JSON report.json machine readable HTML report.html self-contained LEGEND MASKED VALUES NO OUTBOUND SOCKET PIPELINE STAGE CONFIG AND OUTPUT