Public Dataset Endpoint Refresh Pack

$747

one-time

Checkout for this service opens by emailed invoice link. Tell us where to send it and we reply the same working day.

Free specimen and paid-delivery boundary

A completed miniature refresh receipt. It shows the latest adjacent Federal Register pair whose changed digests each identify exactly one retained source row, plus an independently recounted stability finding across the three named clock archives.

Specimen provenance: Captured: 2026-08-26 · Target: one adjacent before/after pair for the exact retained Federal Register newest-documents endpoint · Evidence method: read-only replay of two sealed successful receipts with independent row, response-body, and collection-batch hash checks

Proof scope: method boundary only. This free specimen does not demonstrate the full paid delivery.

What this free specimen actually shows

• One exact adjacent before/after receipt pair for the same URL
• Recomputed byte, content, row, and collection-batch seals
• A CHANGED byte/digest verdict and retained-window stability inventory

Paid outputs not demonstrated by this specimen

• The other two receipts in the paid three-receipt pack
• UNCHANGED receipt branch
• UNKNOWN receipt branch

Paid delivery contract

You send:

• One pre-approved stable public HTTPS endpoint
• Two adjacent retained successful receipts for that exact URL
• The three retained raw-fetch clocks used for the stability population

You get back:

• Prior and new UTC retrieval times, status, byte count, and SHA-256
• A CHANGED verdict with byte and payload-shape comparisons
• The number of comparable exact URLs whose digest stayed stable in-window

Limits: This specimen contains one of the three receipts in the paid pack. CHANGED means only that exact retained response bytes differed; it does not identify a new release, explain why upstream content changed, or redistribute source rows. It does not demonstrate the UNCHANGED or UNKNOWN receipt branches. The stability count covers only successful retained observations and never predicts future behavior.

Open the FDA, Federal Register, and agent-census SQLite stores read-only; recompute every raw-fetch row seal; group exact URLs having at least two successful receipts by distinct body digest; then choose the latest adjacent changed Federal Register pair whose two digests each occur in exactly one source row. For that pair, decompress both retained bodies, recompute byte counts and SHA-256 values, compare JSON shape, and verify each originating collection batch. This is bounded evidence of the dated method stated here; it does not establish affiliation, endorsement, authorization, or any current outcome beyond that scope.