{ "document_id": 259, "filename": "6a782d2c-24d7-45de-8de1-de2f86acc709_France_Server_Architecture_Diagram.pdf", "engine": "paddleocr_vl", "source_run_id": 384, "generated_at": "2026-08-21T08:33:20+00:00", "page_count": 4, "counts_by_type": { "figure": 6, "footer": 9, "header": 3, "heading": 49, "list": 3, "list_item": 12, "table": 4, "table_cell": 27, "text": 34 }, "timings": { "total_ms": 320, "structure_ms": 279, "text_layer_ms": 29, "build_ms": 11 }, "pages": [ { "document_id": 259, "page": 1, "width_pt": 841.92, "height_pt": 594.96, "engine": "paddleocr_vl", "source_run_id": 384, "item_count": 46, "counts_by_type": { "figure": 1, "footer": 3, "header": 1, "heading": 3, "table": 4, "table_cell": 26, "text": 8 }, "build_ms": 5, "items": [ { "id": "p1.1", "order": 1, "type": "header", "bbox": { "l": 29.28, "t": 568.32, "r": 226.56, "b": 557.76 }, "text": "ARCHITECTURE & INTEGRATION MODEL", "attrs": { "block_label": "header" } }, { "id": "p1.2", "order": 2, "type": "heading", "bbox": { "l": 29.52, "t": 549.36, "r": 372.96, "b": 519.84 }, "text": "Split Deployment — France", "attrs": { "block_label": "paragraph_title" } }, { "id": "p1.3", "order": 3, "type": "text", "bbox": { "l": 29.04, "t": 515.04, "r": 540.24, "b": 488.16 }, "text": "Their data plane and our analysis plane both run in France, on separate hardware, joined only by a federated message link.\nPROPOSED" }, { "id": "p1.4", "order": 4, "type": "text", "bbox": { "l": 671.28, "t": 504.72, "r": 803.52, "b": 493.68 }, "text": "PROPOSED | AUGUST 2026" }, { "id": "p1.5", "order": 5, "type": "heading", "bbox": { "l": 42, "t": 446.64, "r": 337.68, "b": 418.08 }, "text": "Their infrastructure\nDATA PLANE · FRANCE · THEIR HARDWARE · THEIR DPA · STATEFUL", "attrs": { "block_label": "paragraph_title" } }, { "id": "p1.6", "order": 6, "type": "text", "bbox": { "l": 42.24, "t": 400.8, "r": 358.32, "b": 358.8 }, "text": "Gateway, 40+ frontends, ~30 data services, MySQL, object store and their own broker. Every finding, screenshot, document, PII analysis and user account lives here.\nPRIVATE LINK\nmTLS\nfederation\nWireGuard\nor IPsec" }, { "id": "p1.7", "order": 7, "type": "table", "bbox": { "l": 45.36, "t": 352.32, "r": 354.96, "b": 291.12 }, "text": "", "attrs": { "block_label": "table", "n_rows": 3, "n_cols": 3, "integrity": { "cells_checked": 7, "cells_flagged": 0 } }, "children": [ { "id": "p1.7.1", "order": 8, "type": "table_cell", "bbox": { "l": 51.749998, "t": 349.16622, "r": 125.999418, "b": 338.44122100000004 }, "text": "Traefik gateway", "attrs": { "row": 0, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true, "column_header": true, "header_source": "heuristic" } }, { "id": "p1.7.2", "order": 9, "type": "table_cell", "bbox": { "l": 143.249994, "t": 349.16622, "r": 202.649539, "b": 338.44122100000004 }, "text": "accounts-api", "attrs": { "row": 0, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true, "column_header": true, "header_source": "heuristic" } }, { "id": "p1.7.3", "order": 10, "type": "table_cell", "bbox": { "l": 219.902338, "t": 349.16622, "r": 284.251842, "b": 338.44122100000004 }, "text": "40+ frontends", "attrs": { "row": 0, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true, "column_header": true, "header_source": "heuristic" } }, { "id": "p1.7.4", "order": 11, "type": "table_cell", "bbox": { "l": 51.749998, "t": 327.416221, "r": 135.899336, "b": 316.69122200000004 }, "text": "~30 data services", "attrs": { "row": 1, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.7.5", "order": 12, "type": "table_cell", "bbox": { "l": 153.152341, "t": 327.416221, "r": 197.70201, "b": 316.69122200000004 }, "text": "MySQL 8.4", "attrs": { "row": 1, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.7.6", "order": 13, "type": "table_cell", "bbox": { "l": 214.957026, "t": 327.416221, "r": 351.008907, "b": 316.69122200000004 }, "text": "object store their broker", "attrs": { "row": 1, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.7.7", "order": 14, "type": "table_cell", "bbox": { "l": 45.36, "t": 311.52, "r": 354.96, "b": 291.12 }, "text": "DETAIL ON PAGE 2", "attrs": { "row": 2, "col": 0, "rowspan": 1, "colspan": 3 } } ] }, { "id": "p1.8", "order": 15, "type": "text", "bbox": { "l": 387.84, "t": 427.44, "r": 453.84, "b": 364.08 }, "text": "PRIVATE LINK\nmTLS\nfederation\nWireGuard\nor IPsec" }, { "id": "p1.9", "order": 16, "type": "text", "bbox": { "l": 396.24, "t": 334.08, "r": 447.6, "b": 315.6 }, "text": "← verdicts crawl jobs" }, { "id": "p1.10", "order": 17, "type": "heading", "bbox": { "l": 480.48, "t": 446.64, "r": 777.36, "b": 418.32 }, "text": "Our infrastructure\nANALYSIS PLANE · FRANCE · OUR HARDWARE · OUR IP · STATELESS", "attrs": { "block_label": "paragraph_title" } }, { "id": "p1.11", "order": 18, "type": "text", "bbox": { "l": 481.68, "t": 401.04, "r": 777.84, "b": 359.28 }, "text": "Broker, crawler fleet, engine-service, ai-facade and Redis. It holds the tests and prompts and nothing else — no findings, no accounts, no documents." }, { "id": "p1.12", "order": 19, "type": "figure", "bbox": { "l": 457.44, "t": 347.28, "r": 465.84, "b": 338.16 }, "text": "", "attrs": { "block_label": "image" } }, { "id": "p1.13", "order": 20, "type": "table", "bbox": { "l": 489.6, "t": 349.16622, "r": 705.36, "b": 337.68 }, "text": "", "attrs": { "n_rows": 1, "n_cols": 3, "integrity": { "cells_checked": 3, "cells_flagged": 0 } }, "children": [ { "id": "p1.13.1", "order": 21, "type": "table_cell", "bbox": { "l": 490.68748, "t": 349.16622, "r": 540.187107, "b": 338.44122100000004 }, "text": "our broker", "attrs": { "row": 0, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.13.2", "order": 22, "type": "table_cell", "bbox": { "l": 557.437477, "t": 349.16622, "r": 631.686897, "b": 338.44122100000004 }, "text": "crawler-manager", "attrs": { "row": 0, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.13.3", "order": 23, "type": "table_cell", "bbox": { "l": 648.937473, "t": 349.16622, "r": 703.387059, "b": 338.44122100000004 }, "text": "crawler × 10", "attrs": { "row": 0, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true } } ] }, { "id": "p1.14", "order": 24, "type": "table", "bbox": { "l": 555.84, "t": 348.96, "r": 792, "b": 337.2 }, "text": "crawler-manager | crawler x10 | engine-service", "attrs": { "ai_retyped_from": "text", "ai_reason": "Item #24 duplicates the chip row already extracted as table cells; it is a row of chips, not standalone text." }, "children": [ { "type": "table_cell", "bbox": { "l": 711.42, "t": 352.81, "r": 799.82, "b": 337.94 }, "text": "engine-service", "attrs": { "source": "ai_vision", "reason": "The 'engine-service' chip in our-infrastructure chip row is not covered by any table_cell (only mentioned loosely in overlapping text item #24)." }, "children": [], "order": 25, "id": "p1.14.1" } ] }, { "id": "p1.15", "order": 26, "type": "text", "bbox": { "l": 489.6, "t": 327.6, "r": 579.36, "b": 315.84 }, "text": "ai-facade Redis" }, { "id": "p1.16", "order": 27, "type": "text", "bbox": { "l": 482.4, "t": 306.24, "r": 569.76, "b": 295.44 }, "text": "DETAIL ON PAGE 3" }, { "id": "p1.17", "order": 28, "type": "table", "bbox": { "l": 38.16, "t": 259.2, "r": 803.76, "b": 120.72 }, "text": "", "attrs": { "block_label": "table", "n_rows": 5, "n_cols": 3, "integrity": { "cells_checked": 15, "cells_flagged": 0 } }, "children": [ { "id": "p1.17.1", "order": 29, "type": "table_cell", "bbox": { "l": 45.749998, "t": 254.666224, "r": 154.154046, "b": 243.94122500000003 }, "text": "WHERE THE LINE FALLS", "attrs": { "row": 0, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true, "column_header": true, "header_source": "heuristic" } }, { "id": "p1.17.2", "order": 30, "type": "table_cell", "bbox": { "l": 168.749993, "t": 254.666224, "r": 282.598991, "b": 243.94122500000003 }, "text": "ANALYSIS PLANE - OURS", "attrs": { "row": 0, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true, "column_header": true, "header_source": "heuristic" } }, { "id": "p1.17.3", "order": 31, "type": "table_cell", "bbox": { "l": 487.68748, "t": 254.666224, "r": 590.646578, "b": 243.94122500000003 }, "text": "DATA PLANE - THEIRS", "attrs": { "row": 0, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true, "column_header": true, "header_source": "heuristic" } }, { "id": "p1.17.4", "order": 32, "type": "table_cell", "bbox": { "l": 45.749998, "t": 225.703776, "r": 71.158529, "b": 213.73857600000002 }, "text": "Holds", "attrs": { "row": 1, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.5", "order": 33, "type": "table_cell", "bbox": { "l": 168.749993, "t": 225.703776, "r": 283.039643, "b": 213.73857600000002 }, "text": "Tests, crawler, AI prompts", "attrs": { "row": 1, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.6", "order": 34, "type": "table_cell", "bbox": { "l": 487.68748, "t": 225.703776, "r": 682.17094, "b": 213.73857600000002 }, "text": "Every result, document, screenshot, account", "attrs": { "row": 1, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.7", "order": 35, "type": "table_cell", "bbox": { "l": 45.749998, "t": 195.703777, "r": 68.779509, "b": 183.73857700000002 }, "text": "State", "attrs": { "row": 2, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.8", "order": 36, "type": "table_cell", "bbox": { "l": 168.749993, "t": 195.703777, "r": 289.240628, "b": 183.73857700000002 }, "text": "None that survives a restart", "attrs": { "row": 2, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.9", "order": 37, "type": "table_cell", "bbox": { "l": 487.68748, "t": 195.703777, "r": 518.214774, "b": 183.73857700000002 }, "text": "All of it", "attrs": { "row": 2, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.10", "order": 38, "type": "table_cell", "bbox": { "l": 45.749998, "t": 165.703778, "r": 104.484039, "b": 153.73857800000002 }, "text": "Reachable by", "attrs": { "row": 3, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.11", "order": 39, "type": "table_cell", "bbox": { "l": 168.749993, "t": 165.703778, "r": 312.591934, "b": 153.73857800000002 }, "text": "Their broker, over the private link", "attrs": { "row": 3, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.12", "order": 40, "type": "table_cell", "bbox": { "l": 487.68748, "t": 165.703778, "r": 590.589099, "b": 153.73857800000002 }, "text": "Their browsers, directly", "attrs": { "row": 3, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.13", "order": 41, "type": "table_cell", "bbox": { "l": 45.749998, "t": 135.70378000000005, "r": 99.745549, "b": 123.73857900000002 }, "text": "If they leave", "attrs": { "row": 4, "col": 0, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.14", "order": 42, "type": "table_cell", "bbox": { "l": 168.749993, "t": 135.70378000000005, "r": 235.001321, "b": 123.73857900000002 }, "text": "We keep the IP", "attrs": { "row": 4, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } }, { "id": "p1.17.15", "order": 43, "type": "table_cell", "bbox": { "l": 487.68748, "t": 135.70378000000005, "r": 577.026819, "b": 123.73857900000002 }, "text": "They keep their data", "attrs": { "row": 4, "col": 2, "rowspan": 1, "colspan": 1, "bbox_refit": true } } ] }, { "id": "p1.18", "order": 44, "type": "footer", "bbox": { "l": 29.04, "t": 36, "r": 263.04, "b": 27.36 }, "text": "SPLIT DEPLOYMENT · FRANCE · ARCHITECTURE TEAM", "attrs": { "block_label": "footer" } }, { "id": "p1.19", "order": 45, "type": "footer", "bbox": { "l": 359.52, "t": 36.24, "r": 685.2, "b": 27.12 }, "text": "BLUE = THEIR HARDWARE, THEIR DPA · AMBER = OUR HARDWARE, OUR IP", "attrs": { "block_label": "footer" } }, { "id": "p1.20", "order": 46, "type": "footer", "bbox": { "l": 782.4, "t": 36.48, "r": 812.16, "b": 26.4 }, "text": "1 / 4", "attrs": { "block_label": "number" } } ], "ai_vision": { "model": "claude-opus-5", "confidence": 78, "added": 1, "retyped": 1 } }, { "document_id": 259, "page": 2, "width_pt": 841.92, "height_pt": 594.96, "engine": "paddleocr_vl", "source_run_id": 384, "item_count": 23, "counts_by_type": { "figure": 2, "footer": 2, "heading": 11, "list": 1, "list_item": 3, "text": 4 }, "build_ms": 1, "items": [ { "id": "p2.1", "order": 1, "type": "figure", "bbox": { "l": 731.04, "t": 566.4, "r": 802.56, "b": 530.16 }, "text": "", "attrs": { "role": "page_header", "block_label": "header_image", "embedded_text": "STATEFUL" } }, { "id": "p2.2", "order": 2, "type": "heading", "bbox": { "l": 44.16, "t": 553.92, "r": 402.24, "b": 537.12 }, "text": "Their infrastructure — what runs on their server", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.3", "order": 3, "type": "heading", "bbox": { "l": 43.92, "t": 533.28, "r": 309.6, "b": 522.24 }, "text": "DATA PLANE · FRANCE · THEIR HARDWARE · THEIR DPA", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4", "order": 4, "type": "figure", "bbox": { "l": 33.84, "t": 495.36, "r": 471.36, "b": 205.68 }, "text": "", "attrs": { "block_label": "image", "embedded_text": "FORWARD-AUTH →\ntheir domain · their TLS certificates accounts-api\nsessions & user accounts\nstatic JS served by nginx\n~30 read/write APIs\nlandscape evidence content seo geo security privacy okf stats\nroadmap academy guardian toolbar feedback accounts …\nTheir broker\nscreenshots · OKF · PDFs RabbitMQ · federated to ours" }, "children": [ { "id": "p2.4.1", "order": 5, "type": "text", "bbox": { "l": 41.28, "t": 490.08, "r": 221.28, "b": 475.92 }, "text": "Their staff browsers, inside their estate", "attrs": { "block_label": "paragraph_title", "ai_retyped_from": "heading", "ai_reason": "Item combines heading 'Their staff' with running text; but primarily it is a diagram node label — the extracted span also mismatches; still it is a labeled row…" } }, { "id": "p2.4.2", "order": 6, "type": "heading", "bbox": { "l": 443.04, "t": 488.4, "r": 471.84, "b": 478.8 }, "text": "HTTPS", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4.3", "order": 7, "type": "heading", "bbox": { "l": 41.76, "t": 446.64, "r": 127.68, "b": 432 }, "text": "Traefik gateway", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4.4", "order": 8, "type": "heading", "bbox": { "l": 40.56, "t": 378.96, "r": 118.56, "b": 364.8 }, "text": "40+ frontends", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4.5", "order": 9, "type": "heading", "bbox": { "l": 41.52, "t": 332.4, "r": 116.4, "b": 318.96 }, "text": "Data services", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4.6", "order": 10, "type": "heading", "bbox": { "l": 39.84, "t": 240.72, "r": 96, "b": 227.04 }, "text": "MySQL 8.4", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4.7", "order": 11, "type": "heading", "bbox": { "l": 188.88, "t": 240.48, "r": 252.24, "b": 227.04 }, "text": "Object store", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.4.8", "order": 12, "type": "text", "bbox": { "l": 39.6, "t": 224.88, "r": 150, "b": 214.08 }, "text": "16 schemas, one per service" } ] }, { "id": "p2.5", "order": 13, "type": "heading", "bbox": { "l": 495.84, "t": 488.4, "r": 615.36, "b": 477.6 }, "text": "EVERYTHING STORED HERE", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.6", "order": 14, "type": "text", "bbox": { "l": 494.64, "t": 471.36, "r": 798, "b": 430.32 }, "text": "Findings, screenshots, documents, PII analysis, remediation tracking and every user account — on infrastructure they own, under their own DPA." }, { "id": "p2.7", "order": 15, "type": "heading", "bbox": { "l": 495.84, "t": 396.96, "r": 593.04, "b": 383.04 }, "text": "What they control", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.8", "order": 16, "type": "list", "bbox": { "l": 495.36, "t": 376.56, "r": 784.56, "b": 294.24 }, "text": "", "attrs": { "item_count": 3 }, "children": [ { "id": "p2.8.1", "order": 17, "type": "list_item", "bbox": { "l": 495.6, "t": 376.56, "r": 750.48, "b": 363.36 }, "text": "Database backups, access controls and disaster recovery." }, { "id": "p2.8.2", "order": 18, "type": "list_item", "bbox": { "l": 495.36, "t": 356.16, "r": 784.56, "b": 328.8 }, "text": "No analytics and no telemetry leave this estate — we cannot read their results." }, { "id": "p2.8.3", "order": 19, "type": "list_item", "bbox": { "l": 495.36, "t": 321.12, "r": 774.48, "b": 294.24 }, "text": "Single-jurisdiction infrastructure: no multi-region replication to reason about." } ] }, { "id": "p2.9", "order": 20, "type": "heading", "bbox": { "l": 496.32, "t": 262.08, "r": 589.44, "b": 251.04 }, "text": "HOW THEY REACH US", "attrs": { "block_label": "paragraph_title" } }, { "id": "p2.10", "order": 21, "type": "text", "bbox": { "l": 494.16, "t": 246.96, "r": 794.16, "b": 218.88 }, "text": "They never call us over HTTP. Their frontends and services talk only to their own broker; the broker federates across the private link." }, { "id": "p2.11", "order": 22, "type": "footer", "bbox": { "l": 29.28, "t": 36, "r": 160.56, "b": 27.12 }, "text": "SPLIT DEPLOYMENT · FRANCE", "attrs": { "block_label": "footer" } }, { "id": "p2.12", "order": 23, "type": "footer", "bbox": { "l": 782.64, "t": 36.24, "r": 812.16, "b": 26.4 }, "text": "2 / 4", "attrs": { "block_label": "number" } } ], "ai_vision": { "model": "claude-opus-5", "confidence": 76, "added": 0, "retyped": 1 } }, { "document_id": 259, "page": 3, "width_pt": 841.92, "height_pt": 594.96, "engine": "paddleocr_vl", "source_run_id": 384, "item_count": 32, "counts_by_type": { "footer": 2, "header": 1, "heading": 12, "table_cell": 2, "text": 15 }, "build_ms": 2, "items": [ { "id": "p3.1", "order": 1, "type": "header", "bbox": { "l": 732.24, "t": 545.52, "r": 788.16, "b": 534.72 }, "text": "STATELESS", "attrs": { "block_label": "header" } }, { "id": "p3.2", "order": 2, "type": "heading", "bbox": { "l": 44.4, "t": 553.44, "r": 380.64, "b": 537.12 }, "text": "Our infrastructure — what runs on our server", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.3", "order": 3, "type": "heading", "bbox": { "l": 43.68, "t": 533.52, "r": 303.84, "b": 522.24 }, "text": "ANALYSIS PLANE · FRANCE · OUR HARDWARE · OUR IP", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.4", "order": 4, "type": "heading", "bbox": { "l": 43.2, "t": 487.44, "r": 163.44, "b": 473.52 }, "text": "Our RabbitMQ broker", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.5", "order": 5, "type": "text", "bbox": { "l": 42.96, "t": 471.6, "r": 197.52, "b": 460.8 }, "text": "publishes results out · never stores them" }, { "id": "p3.6", "order": 6, "type": "heading", "bbox": { "l": 400.8, "t": 480.24, "r": 451.68, "b": 469.2 }, "text": "FEDERATED", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.7", "order": 7, "type": "text", "bbox": { "l": 40.56, "t": 429.12, "r": 326.4, "b": 415.92 }, "text": "", "attrs": { "n_rows": 1, "n_cols": 2, "integrity": { "cells_checked": 1, "cells_flagged": 0 }, "ai_retyped_from": "table", "ai_reason": "The region is two side-by-side cards (crawler-manager / crawler ×10), not a table with cells." }, "children": [ { "id": "p3.7.1", "order": 8, "type": "table_cell", "bbox": { "l": 40.56, "t": 429.12, "r": 183.48, "b": 415.92 }, "text": "", "attrs": { "row": 0, "col": 0, "rowspan": 1, "colspan": 1 } }, { "id": "p3.7.2", "order": 9, "type": "table_cell", "bbox": { "l": 265.910149, "t": 428.876367, "r": 324.172698, "b": 415.990767 }, "text": "crawler × 10", "attrs": { "row": 0, "col": 1, "rowspan": 1, "colspan": 1, "bbox_refit": true } } ] }, { "id": "p3.8", "order": 10, "type": "text", "bbox": { "l": 40.8, "t": 413.76, "r": 378.24, "b": 402.48 }, "text": "assigns crawl slots renders pages · runs 6 passes" }, { "id": "p3.9", "order": 11, "type": "heading", "bbox": { "l": 41.28, "t": 377.52, "r": 116.64, "b": 364.8 }, "text": "engine-service", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.10", "order": 12, "type": "heading", "bbox": { "l": 264.24, "t": 377.52, "r": 313.92, "b": 365.28 }, "text": "ai-facade", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.11", "order": 13, "type": "text", "bbox": { "l": 41.04, "t": 362.16, "r": 345.12, "b": 350.64 }, "text": "single-URL scans holds the prompt set" }, { "id": "p3.12", "order": 14, "type": "heading", "bbox": { "l": 496.8, "t": 488.16, "r": 599.76, "b": 477.6 }, "text": "NOTHING STORED HERE", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.13", "order": 15, "type": "text", "bbox": { "l": 494.4, "t": 471.36, "r": 792.48, "b": 430.32 }, "text": "No findings, no accounts, no documents. Their data never travels to or resides on our infrastructure — which is why no code disclosure is needed." }, { "id": "p3.14", "order": 16, "type": "text", "bbox": { "l": 496.08, "t": 398.64, "r": 571.44, "b": 387.6 }, "text": "OUTBOUND HTTPS" }, { "id": "p3.15", "order": 17, "type": "heading", "bbox": { "l": 496.32, "t": 384.48, "r": 604.08, "b": 371.28 }, "text": "Their public websites", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.16", "order": 18, "type": "text", "bbox": { "l": 495.12, "t": 368.16, "r": 682.08, "b": 355.68 }, "text": "the pages under test, fetched by the crawlers" }, { "id": "p3.17", "order": 19, "type": "heading", "bbox": { "l": 41.04, "t": 321.12, "r": 72, "b": 308.88 }, "text": "Redis", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.18", "order": 20, "type": "text", "bbox": { "l": 240.72, "t": 319.44, "r": 461.52, "b": 308.64 }, "text": "frontier · seen-set · slot queue — nothing survives a restart" }, { "id": "p3.19", "order": 21, "type": "heading", "bbox": { "l": 43.44, "t": 275.76, "r": 174.72, "b": 264.48 }, "text": "★ NEVER LEAVES THIS BOX", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.20", "order": 22, "type": "text", "bbox": { "l": 48.48, "t": 254.88, "r": 370.8, "b": 242.16 }, "text": "content analyzer\nSEO / GEO analyzers" }, { "id": "p3.21", "order": 23, "type": "text", "bbox": { "l": 49.68, "t": 254.64, "r": 158.4, "b": 242.64 }, "text": "86 WCAG SC test files" }, { "id": "p3.22", "order": 24, "type": "text", "bbox": { "l": 48.48, "t": 231.36, "r": 410.16, "b": 219.12 }, "text": "PDF remediation engine" }, { "id": "p3.23", "order": 25, "type": "text", "bbox": { "l": 49.92, "t": 230.88, "r": 281.52, "b": 219.36 }, "text": "web security analyzer data privacy analyzer" }, { "id": "p3.24", "order": 26, "type": "text", "bbox": { "l": 50.16, "t": 208.32, "r": 119.28, "b": 196.56 }, "text": "AI prompt set" }, { "id": "p3.25", "order": 27, "type": "text", "bbox": { "l": 42, "t": 185.28, "r": 426, "b": 159.6 }, "text": "Not compiled into their binaries, not shipped as an image they can inspect, not present in their database. They receive verdicts only." }, { "id": "p3.26", "order": 28, "type": "heading", "bbox": { "l": 496.32, "t": 326.64, "r": 581.28, "b": 315.6 }, "text": "OUTBOUND HTTPS ▲", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.27", "order": 29, "type": "heading", "bbox": { "l": 496.08, "t": 312.24, "r": 537.84, "b": 299.28 }, "text": "OpenAI", "attrs": { "block_label": "paragraph_title" } }, { "id": "p3.28", "order": 30, "type": "text", "bbox": { "l": 494.88, "t": 295.92, "r": 790.32, "b": 270.72 }, "text": "Calls may include page content. Their DPO should review OpenAI's DPA, or AI features can be disabled." }, { "id": "p3.29", "order": 31, "type": "footer", "bbox": { "l": 29.28, "t": 36, "r": 160.56, "b": 27.12 }, "text": "SPLIT DEPLOYMENT · FRANCE", "attrs": { "block_label": "footer" } }, { "id": "p3.30", "order": 32, "type": "footer", "bbox": { "l": 783.36, "t": 36.24, "r": 812.16, "b": 26.4 }, "text": "3 / 4", "attrs": { "block_label": "number" } } ], "ai_vision": { "model": "claude-opus-5", "confidence": 62, "added": 0, "retyped": 1 } }, { "document_id": 259, "page": 4, "width_pt": 841.92, "height_pt": 594.96, "engine": "paddleocr_vl", "source_run_id": 384, "item_count": 47, "counts_by_type": { "footer": 2, "header": 1, "heading": 22, "list": 2, "list_item": 9, "text": 11 }, "build_ms": 3, "items": [ { "id": "p4.1", "order": 1, "type": "header", "bbox": { "l": 218.16, "t": 558.96, "r": 513.84, "b": 545.76 }, "text": "every step stays in France; the only crossing is the federated queue", "attrs": { "block_label": "header" } }, { "id": "p4.2", "order": 2, "type": "heading", "bbox": { "l": 30.24, "t": 564.72, "r": 210.72, "b": 545.28 }, "text": "One crawl, end to end", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.3", "order": 3, "type": "heading", "bbox": { "l": 40.08, "t": 506.64, "r": 59.52, "b": 492.48 }, "text": "01", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.4", "order": 4, "type": "heading", "bbox": { "l": 178.32, "t": 504.24, "r": 206.88, "b": 494.16 }, "text": "THEIRS", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.5", "order": 5, "type": "heading", "bbox": { "l": 237.36, "t": 506.4, "r": 256.08, "b": 492.24 }, "text": "02", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.6", "order": 6, "type": "heading", "bbox": { "l": 374.64, "t": 504.24, "r": 404.4, "b": 494.16 }, "text": "THEIRS", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.7", "order": 7, "type": "heading", "bbox": { "l": 39.84, "t": 486, "r": 119.28, "b": 473.28 }, "text": "Scan requested", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.8", "order": 8, "type": "heading", "bbox": { "l": 237.12, "t": 486, "r": 341.28, "b": 473.28 }, "text": "Job onto their queue", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.9", "order": 9, "type": "text", "bbox": { "l": 39.6, "t": 468, "r": 409.68, "b": 443.76 }, "text": "A user in their UI asks for a crawl; the data service records the job.\nThe request is published to their broker — no HTTP call to us." }, { "id": "p4.10", "order": 10, "type": "heading", "bbox": { "l": 433.92, "t": 506.64, "r": 452.88, "b": 492.24 }, "text": "03", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.11", "order": 11, "type": "text", "bbox": { "l": 576.48, "t": 505.2, "r": 602.64, "b": 492.96 }, "text": "", "attrs": { "block_label": "image", "embedded_text": "LINK", "ai_retyped_from": "figure", "ai_reason": "This is the 'LINK' badge label text on card 03, not a figure." } }, { "id": "p4.12", "order": 12, "type": "heading", "bbox": { "l": 433.44, "t": 485.76, "r": 521.28, "b": 474 }, "text": "Federated across", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.13", "order": 13, "type": "text", "bbox": { "l": 433.2, "t": 468.24, "r": 597.36, "b": 444.48 }, "text": "The job crosses the mTLS private link to our broker." }, { "id": "p4.14", "order": 14, "type": "heading", "bbox": { "l": 630.72, "t": 506.4, "r": 649.44, "b": 492.48 }, "text": "04", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.15", "order": 15, "type": "heading", "bbox": { "l": 774.72, "t": 504.24, "r": 797.76, "b": 493.92 }, "text": "OURS", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.16", "order": 16, "type": "heading", "bbox": { "l": 630.24, "t": 486.24, "r": 703.68, "b": 473.04 }, "text": "Slots assigned", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.17", "order": 17, "type": "text", "bbox": { "l": 630, "t": 468, "r": 789.6, "b": 444 }, "text": "crawler-manager claims a slot in Redis and dispatches crawlers." }, { "id": "p4.18", "order": 18, "type": "heading", "bbox": { "l": 40.32, "t": 405.36, "r": 59.52, "b": 391.2 }, "text": "05", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.19", "order": 19, "type": "text", "bbox": { "l": 183.6, "t": 403.44, "r": 208.32, "b": 392.4 }, "text": "", "attrs": { "block_label": "image", "embedded_text": "OURS", "ai_retyped_from": "figure", "ai_reason": "This is the 'OURS' badge label text on card 05, not a figure." } }, { "id": "p4.20", "order": 20, "type": "heading", "bbox": { "l": 39.84, "t": 384.72, "r": 114.24, "b": 372.24 }, "text": "Six passes run", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.21", "order": 21, "type": "text", "bbox": { "l": 39.36, "t": 366.72, "r": 203.28, "b": 329.52 }, "text": "Crawlers fetch their public pages and run WCAG, Content, SEO, GEO, Security and Privacy." }, { "id": "p4.22", "order": 22, "type": "heading", "bbox": { "l": 237.36, "t": 405.12, "r": 255.6, "b": 390.96 }, "text": "06", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.23", "order": 23, "type": "text", "bbox": { "l": 380.16, "t": 403.44, "r": 406.08, "b": 392.16 }, "text": "", "attrs": { "block_label": "image", "embedded_text": "LINK", "ai_retyped_from": "figure", "ai_reason": "This is the 'LINK' badge label text on card 06, not a figure." } }, { "id": "p4.24", "order": 24, "type": "heading", "bbox": { "l": 236.88, "t": 384.96, "r": 330.96, "b": 371.76 }, "text": "Verdicts published", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.25", "order": 25, "type": "text", "bbox": { "l": 235.68, "t": 366.72, "r": 406.08, "b": 342.72 }, "text": "Structured results — not rules — federate back to their broker." }, { "id": "p4.26", "order": 26, "type": "heading", "bbox": { "l": 433.92, "t": 405.36, "r": 452.64, "b": 390.96 }, "text": "07", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.27", "order": 27, "type": "heading", "bbox": { "l": 570.24, "t": 402.72, "r": 601.2, "b": 393.36 }, "text": "THEIRS", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.28", "order": 28, "type": "heading", "bbox": { "l": 433.44, "t": 384.96, "r": 524.88, "b": 372.24 }, "text": "Stored and shown", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.29", "order": 29, "type": "text", "bbox": { "l": 433.44, "t": 366.72, "r": 593.76, "b": 342.48 }, "text": "Their data services write to MySQL and the object store; the UI updates." }, { "id": "p4.30", "order": 30, "type": "heading", "bbox": { "l": 630.96, "t": 408.24, "r": 682.32, "b": 399.36 }, "text": "NET EFFECT", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.31", "order": 31, "type": "text", "bbox": { "l": 630.72, "t": 395.04, "r": 769.2, "b": 382.8 }, "text": "Verdicts move, data doesn't" }, { "id": "p4.32", "order": 32, "type": "text", "bbox": { "l": 630.72, "t": 376.8, "r": 791.28, "b": 352.56 }, "text": "They hold every byte of customer data; we hold every line of analysis IP." }, { "id": "p4.33", "order": 33, "type": "heading", "bbox": { "l": 43.68, "t": 291.36, "r": 146.16, "b": 275.76 }, "text": "Why this holds up", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.34", "order": 34, "type": "list", "bbox": { "l": 42.24, "t": 267.6, "r": 396.96, "b": 143.52 }, "text": "", "attrs": { "item_count": 5 }, "children": [ { "id": "p4.34.1", "order": 35, "type": "list_item", "bbox": { "l": 42.24, "t": 267.6, "r": 396.96, "b": 254.4 }, "text": "No shared secrets — the link authenticates with mTLS certificates, not API keys." }, { "id": "p4.34.2", "order": 36, "type": "list_item", "bbox": { "l": 42.96, "t": 246.72, "r": 375.6, "b": 219.12 }, "text": "No API exposure — our analysis services answer no HTTP; the queue is the only door." }, { "id": "p4.34.3", "order": 37, "type": "list_item", "bbox": { "l": 42.96, "t": 211.44, "r": 360.24, "b": 198.24 }, "text": "Outage resilient — if the link drops, both queues buffer until it recovers." }, { "id": "p4.34.4", "order": 38, "type": "list_item", "bbox": { "l": 42.96, "t": 191.52, "r": 379.92, "b": 178.32 }, "text": "Independent audit — both brokers log federation, reviewable on either side." }, { "id": "p4.34.5", "order": 39, "type": "list_item", "bbox": { "l": 43.2, "t": 170.16, "r": 383.04, "b": 143.52 }, "text": "One exception — ai-facade calls OpenAI and may include page content; their DPO decides." } ] }, { "id": "p4.35", "order": 40, "type": "heading", "bbox": { "l": 440.16, "t": 291.12, "r": 513.12, "b": 276.24 }, "text": "If they leave", "attrs": { "block_label": "paragraph_title" } }, { "id": "p4.36", "order": 41, "type": "list", "bbox": { "l": 438.96, "t": 267.6, "r": 789.84, "b": 163.68 }, "text": "", "attrs": { "item_count": 4 }, "children": [ { "id": "p4.36.1", "order": 42, "type": "list_item", "bbox": { "l": 438.96, "t": 267.6, "r": 755.52, "b": 240.48 }, "text": "They export findings, evidence and remediation tracking from their own data services." }, { "id": "p4.36.2", "order": 43, "type": "list_item", "bbox": { "l": 439.2, "t": 232.8, "r": 765.36, "b": 219.12 }, "text": "We stop publishing to their queue and the private link is decommissioned." }, { "id": "p4.36.3", "order": 44, "type": "list_item", "bbox": { "l": 439.44, "t": 211.44, "r": 752.16, "b": 185.28 }, "text": "They keep every finding, screenshot, document and account – no data escrow needed." }, { "id": "p4.36.4", "order": 45, "type": "list_item", "bbox": { "l": 439.2, "t": 177.36, "r": 789.84, "b": 163.68 }, "text": "We retain the full analysis IP; no code was ever disclosed, so none is recovered." } ] }, { "id": "p4.37", "order": 46, "type": "footer", "bbox": { "l": 29.52, "t": 36, "r": 262.8, "b": 27.36 }, "text": "SPLIT DEPLOYMENT · FRANCE · ARCHITECTURE TEAM", "attrs": { "block_label": "footer" } }, { "id": "p4.38", "order": 47, "type": "footer", "bbox": { "l": 783.12, "t": 36.24, "r": 812.16, "b": 26.4 }, "text": "4 / 4", "attrs": { "block_label": "number" } } ], "ai_vision": { "model": "claude-opus-5", "confidence": 78, "added": 0, "retyped": 3 } } ] }