Document organization example · synthetic data
Inspect the organized files, not just a product description.
Two mixed deliveries, a reviewed account-and-period inventory, and the actual separated source pages. Download the result of an isolated local, reviewer-assisted workflow.
All people, institutions and records are fictional. Native-text proposals were followed by scripted synthetic review actions, not independent practitioner approval. This is not a recognition-accuracy or production-performance benchmark.
No account needed. PDFs contain selectable text. If your browser cannot preview a PDF, download it and open it in a PDF reader. These files are separate from the SIM-001 read-only workspace.
What arrived; what remains to review
- Delivered PDFs / pages
- 4 / 57
- Delivery batches
- 2
- Distinct periods represented
- 17
- Retained document occurrences
- 19
- Duplicate occurrences included
- 2
- Outstanding / incomplete / unresolved page
- 1 / 1 / 1
| Account | 2025-01 | 2025-02 | 2025-03 | 2025-04 | 2025-05 | 2025-06 |
|---|---|---|---|---|---|---|
| Alder Example Bank Household checking · •••• 1842 | Present | Present | Present | Present | Present | Present |
| Birch Example Bank Reserve savings · •••• 2953 | Present | Present | Present | Present | Page gap | Present |
| Cedar Example Card Household card · •••• 3064 | Present | Present | Present | Present | Present | Open |
What to request next
- Cedar Example Card · •••• 3064: 2025-06-01 to 2025-06-30 is expected but not represented.
- D13 — Birch Example Bank · •••• 2953: INCOMPLETE: printed pages 1, 3; missing printed page(s) 2. Approve only the received range and visible period metadata; request missing page. Source: download_2.pdf, pp. 10–11.
- The ambiguous continuation page remains unresolved in its original delivery. See the workpaper and ZIP's unresolved-items file; it is not an approved separated document.
Original inputs and reviewed inventory
Original delivery PDFs are preserved unchanged in Originals/. Reviewed segments contain the original source pages, not statements reconstructed from extracted text.
- scan_004.pdf — 27 pages
- doc.pdf — 3 pages
- download_2.pdf — 24 pages
- new_final.pdf — 3 pages
| Reference | Institution / account | Period | Delivery / original pages | Review status |
|---|---|---|---|---|
| D01 | Alder Example Bank •••• 1842 | 2025-01-01 2025-01-31 | Batch 1 doc.pdf pp. 1–3 | Approved range / metadata |
| D02 | Alder Example Bank •••• 1842 | 2025-01-01 2025-01-31 | Batch 2 new_final.pdf pp. 1–3 | Approved range / metadata; duplicate copy |
| D03 | Alder Example Bank •••• 1842 | 2025-02-01 2025-02-28 | Batch 1 scan_004.pdf pp. 4–6 | Approved range / metadata |
| D04 | Alder Example Bank •••• 1842 | 2025-03-01 2025-03-31 | Batch 1 scan_004.pdf pp. 19–21 | Approved range / metadata |
| D05 | Alder Example Bank •••• 1842 | 2025-04-01 2025-04-30 | Batch 2 download_2.pdf pp. 4–6 | Approved range / metadata |
| D06 | Alder Example Bank •••• 1842 | 2025-05-01 2025-05-31 | Batch 1 scan_004.pdf pp. 13–15 | Approved range / metadata |
| D07 | Alder Example Bank •••• 1842 | 2025-06-01 2025-06-30 | Batch 2 download_2.pdf pp. 12–14 | Approved range / metadata |
| D08 | Birch Example Bank •••• 2953 | 2025-01-01 2025-01-31 | Batch 1 scan_004.pdf pp. 10–12 | Approved range / metadata |
| D09 | Birch Example Bank •••• 2953 | 2025-02-01 2025-02-28 | Batch 1 scan_004.pdf pp. 22–24 | Approved range / metadata |
| D10 | Birch Example Bank •••• 2953 | 2025-02-01 2025-02-28 | Batch 2 download_2.pdf pp. 21–23 | Approved range / metadata; duplicate copy |
| D11 | Birch Example Bank •••• 2953 | 2025-03-01 2025-03-31 | Batch 1 scan_004.pdf pp. 1–3 | Approved range / metadata |
| D12 | Birch Example Bank •••• 2953 | 2025-04-01 2025-04-30 | Batch 2 download_2.pdf pp. 18–20 | Approved range / metadata |
| D13 | Birch Example Bank •••• 2953 | 2025-05-01 2025-05-31 | Batch 2 download_2.pdf pp. 10–11 | Approved received pages; INCOMPLETE |
| D14 | Birch Example Bank •••• 2953 | 2025-06-01 2025-06-30 | Batch 2 download_2.pdf pp. 1–3 | Approved range / metadata |
| D15 | Cedar Example Card •••• 3064 | 2025-01-01 2025-01-31 | Batch 1 scan_004.pdf pp. 7–9 | Approved range / metadata |
| D16 | Cedar Example Card •••• 3064 | 2025-02-01 2025-02-28 | Batch 1 scan_004.pdf pp. 25–27 | Approved range / metadata |
| D17 | Cedar Example Card •••• 3064 | 2025-03-01 2025-03-31 | Batch 1 scan_004.pdf pp. 16–18 | Approved range / metadata |
| D18 | Cedar Example Card •••• 3064 | 2025-04-01 2025-04-30 | Batch 2 download_2.pdf pp. 15–17 | Approved range / metadata |
| D19 | Cedar Example Card •••• 3064 | 2025-05-01 2025-05-31 | Batch 2 download_2.pdf pp. 7–9 | Approved range / metadata |
Consolidated duplicate review
D01 retained for organization; D02 preserved as a duplicate occurrence. Comparison methods: authoritative-plaintext-sha256-and-page-range. Synthetic scripted review: compared complete received page text and sequence; matching statement copy. Retain earlier delivery for coverage; preserve both occurrences. This is not an authenticity finding.
D09 retained for organization; D10 preserved as a duplicate occurrence. Comparison methods: approved-metadata-identity, normalized-page-text-sha256. Synthetic scripted review: compared complete received page text and sequence; matching statement copy. Retain earlier delivery for coverage; preserve both occurrences. This is not an authenticity finding.
The ZIP also contains inventory.csv, manifest.json, original-page mappings, per-file SHA-256 hashes, proposal records and individual review events. Hashes identify exact bytes against recorded references; they do not establish authenticity.
How this example was produced
The local synthetic-reviewer actor recorded scripted review actions through the same revision-checked review services used by the application. The review aid reads visible institution/account/period headers and printed page sequences from ingested page text. It never reads the ground-truth manifest. These actions demonstrate reviewer-assisted organization; they are not independent practitioner review or an automated recognition accuracy claim. All metadata approvals are review actions.
- Native-text PDFs only in this sample. No scanned-image/OCR or Azure recognition performance was tested. No external provider was called.
- Internal missing-page identification is a recorded review observation, not a native coverage-engine result. Represented dates do not imply complete pages or discovery sufficiency.
- The ambiguous continuation page stays in its original delivery; unapproved material is not presented as an approved separated document.
- Two duplicate occurrences are preserved. Multiple native comparison signals can describe the same pair; signals are not extra copies.
- No transaction extraction, ledger reconciliation, LIBR, professional conclusions or admissibility claims are included. Source statement arithmetic is fixture data.
- SHA-256 hashes verify exact bytes against recorded references, not authenticity. SQLite/eager tasks demonstrate local workflow, not production concurrency or scale.
| Input / workflow | What to expect |
|---|---|
| Native-text mixed PDFs | Reviewable document-boundary and metadata proposals for supported patterns. Human approval is required; uncertain labels remain for review. |
| Image-only production pages | Image-only production pages remain unresolved in this workflow. Manual review is needed; document organization does not run OCR. |
| Separate transaction extraction | Configured OCR may apply through the separate financial extraction path, with its own provider controls. Extracted candidates are not automatically approved ledger records. |