logologo
  • How it works
  • Why It Matters
  • Statistics
  • Pricing
  • API
logologo
  • How it works
  • Why It Matters
  • Statistics
  • Pricing
  • API
HTPBE?

Structural PDF tamper detection API. Catches edits your KYC stack misses.

Product

  • How It Works
  • Why It Matters
  • Use Cases
  • Pricing

Developers

  • API Reference
  • GitHub/docs
  • Changelogv2.23.1

Resources

  • FAQ
  • Blog
  • Comparisons
  • Legal & Imprint

© 2024–2026 TMI Iurii Rogulia · VAT ID: FI29845875 · Made in Finland 🇫🇮

Status

Algorithm v2.23.1

Tool profile

Safari

Safari appears on both legitimate first-generation output and downstream re-save flows — context (the other tool on the same document) is what flips the signal.

Back to all statistics
Forensic verdict

Mixed signal

Based on 6 appearances across the HTPBE? corpus.

Modification rate
50%+2pp above baseline
Corpus baseline: 48%
Total appearances
6
0.10% of corpus
Modification rate
50%
+2pp above baseline
Role split
100%C/0%P
Creator vs Producer share of appearances

Corpus profile

How Safari shows up in HTPBE? corpus

Safari is one of the PDF-handling tools surfaced in the HTPBE? corpus. Safari appears predominantly as the original Creator (100% of its occurrences) — i.e. on documents that started life inside Safari rather than passing through it as a downstream re-saver.

In the HTPBE? corpus the contextual signal we look for is a producer/creator mismatch: when Safari appears as the latest Producer on a document whose Creator was an institutional source (e.g. Adobe PDF Library, Microsoft Word, a banking back-end), the document was rebuilt or re-saved after its original creation. That mismatch is the marker — never the tool itself.

On documents where Safari acts as Creator, 50% carry modification markers; on documents where it acts as Producer, 0% do. These are observed rates inside the HTPBE? corpus and should be read as base-rates, not as accusations against Safari or its users.

The signal
In the HTPBE? corpus the contextual signal we look for is a producer/creator mismatch: when Safari appears as the latest Producer on a document whose Creator was an institutional source (e.g. Adobe PDF Library, Microsoft Word, a banking back-end), the document was rebuilt or re-saved after its original creation. That mismatch is the marker — never the tool itself.

Role in the workflow

How Safari shows up in metadata

Every PDF carries a Creator (the application that produced the original document) and a Producer (the engine that wrote the PDF). The same tool can appear in either slot, with very different modification profiles.

CAs Creator · 100%
As Producer · 0%P
CAs Creator
  • Usage
    6
  • Modification rate
    50%
  • Avg file size
    188 KB
PAs Producer
  • Usage
    0
  • Modification rate
    0%

How to read this

The Creator slot typically reflects where a document started life. The Producer slot reflects whatever wrote the bytes — and is the field that gets overwritten when a PDF is opened, edited, and saved by a downstream tool.

A higher modification rate as Producer than as Creator usually means the tool is acting as a re-saver on documents that originated elsewhere. A higher rate as Creator points to fragile workflows around the original authoring app.

Distributions

What ships alongside Safari

The PDF versions Safari writes when acting as Producer, and the other tools that appear in the same documents.

Common Producers when Safari is the Creator

iOS Print to PDF writes 50% of these files — that pairing is the Adobe-stack default for many institutional pipelines.

iOS Print to PDF50.0%
GPL Ghostscript33.3%
macOS Print to PDF16.7%

Related profiles

Tools you’ll see next to Safari

Other tools that frequently share metadata with Safari in the same documents. Each card links to its own forensic profile.

P50% co-occurrence
iOS Print to PDF
Appearances347
Mod rate68%
P33% co-occurrence
GPL Ghostscript
Appearances35
Mod rate43%
P17% co-occurrence
macOS Print to PDF
Appearances747
Mod rate64%

Long tail

Notable observations

Smaller cuts of the Safari corpus — useful context, but treat each row as a single data point rather than a strong signal.

Pages parsed
7
Oldest observed
5 Feb 2026 — 4 months ago

Secure your workflow

Create your account — API key on signup, free test environment on every plan.
From $15/mo. No sales call. Cancel any time.

Start free — close the structural fraud gapSee pricing
Read API docs →