Mobile Forensics · Graduation Project · BCIT 2026

Snapchat Forensics

By Austin Webber — Co-authored by Kaiden Koran

Type Academic Research
Duration Jan – Feb 2026
Platforms Android & iOS
Course BCIT DFCS — FSCT 8590
Android iOS QUIC / TLS 1.3 Magnet AXIOM Cellebrite FTK Imager Frida mitmproxy ADB Wireshark

Overview

Who: A two-person research team at BCIT — Austin Webber and Kaiden Koran — as part of the FSCT 8590 Network Exploits and Vulnerabilities course, with test devices operated under fictional personas (John Titor, Ryoma Echizen, Dean Winchester, Sam Winchester).

What: A multi-vector forensic investigation into Snapchat, targeting the recovery of disappearing messages, media artefacts, and account metadata across both Android and iOS platforms.

When: Controlled study window ran from January 22 to February 19, 2026.

Where: Conducted in an isolated BCIT lab environment. All devices were test accounts on a segregated network with no connection to authentic user data or Snapchat's production infrastructure.

Why: Snapchat is built around ephemerality — messages vanish by design. This project examined whether forensic techniques can recover meaningful evidence despite that design, and where the ceiling of each acquisition method falls.

How: Three simultaneous approaches were pursued: (1) network-layer MITM interception of QUIC/TLS 1.3 traffic; (2) full file system and logical extraction on Android devices; (3) iTunes backup analysis and Keychain extraction on iOS devices. All findings were processed through Magnet AXIOM and Cellebrite Physical Analyzer.

Research Compliance
Conducted entirely with test accounts on isolated lab infrastructure. No third-party data was accessed. Methodology followed ACPO digital evidence principles and Canada's PIPEDA framework.

Lab Setup

The MITM environment was built on a dedicated Linux host acting as a wireless access point. All test devices connected through this host, routing their traffic through mitmproxy for interception. SSL pinning — Snapchat's primary defence against traffic inspection — was bypassed on Android devices using Frida runtime instrumentation injecting a custom script to disable certificate validation at runtime.

Figure 1 — MITM Lab Topology Figure 1 — MITM Lab Topology

Test devices:

  • ASamsung Galaxy A8 (rooted) — "Ryoma Echizen"
  • BSamsung Galaxy S8 (non-rooted) — "John Titor"
  • CiPhone XS (iOS backup) — "Sam Winchester"
  • DiPhone 13 (iOS backup) — "Dean Winchester"

Packet captures (PCAP) were collected using Wireshark on the Linux host over the duration of the study window, capturing all Snapchat-bound traffic for later analysis.

Network Analysis

Snapchat uses QUIC over UDP with TLS 1.3 for all media and message traffic. The combination is effectively opaque to a MITM observer: even with Frida bypassing Android's SSL pinning, Snapchat's additional application-layer certificate pinning (Fidelius) prevented content decryption at the network level.

Key Finding — Network Layer
Message content was not recoverable from network captures. However, the PCAP provided meaningful forensic metadata: server IPs, DNS queries, packet sizes, session timestamps, and the precise moment the Galaxy A8 initiated its first Snapchat session.

This metadata is not trivial — it establishes connection timelines and can corroborate device-level evidence when cross-referenced with SQLite artefacts recovered from storage. It cannot, however, stand alone as evidence of content.

Device Extractions

Android — FFS vs. Logical

The access level of an Android extraction is the single most decisive factor in Snapchat artefact recovery. On the rooted Galaxy A8, a Full File System (FFS) extraction via ADB and Magnet AXIOM produced near-complete results: the Snapchat artefact database (arroyo.db), cached media fragments, account metadata, and partially recoverable deleted message records were all accessible.

On the non-rooted Galaxy S8, logical extraction yielded essentially nothing actionable — Snapchat's sandbox prevents logical tools from reading its data directory without elevated privileges. Additionally, OEM Unlocking was unavailable on this Canadian variant.

Artefact FFS (Rooted A8) Logical (S8)
Chat records (arroyo.db) Recovered Not accessible
Cached media Recovered Not accessible
Account metadata Recovered Not accessible
Deleted records Recovered Not accessible

iOS — Backups and Keychain

iTunes encrypted backups were analysed for both iPhones using Magnet AXIOM and Cellebrite Physical Analyzer. Snapchat stores minimal data in standard backups by design — chat content and media are excluded — but several high-value artefact categories were recovered.

  • 1InteractionC database — conversation timeline records, first-contact events, and interaction counts across the study window.
  • 2Keychain artefacts — account activation timestamps (fideliusTransferableDeviceGraph), session tokens (SCOneTapLoginKeychainKey), establishing precise account chronology.
  • 3EXIF location data — a screenshot taken on the iPhone 13 at 3:36 AM on February 13 contained GPS coordinates placing the device in Maple Ridge, BC.
FFS vs Backup — iOS
iTunes backups provide metadata and timeline evidence, but chat content, video snaps, voice notes, and My Eyes Only material require a full file system extraction (via GrayKey or equivalent) to access. The backup is a useful but incomplete source.
Figure 2 — InteractionC Timeline Visualisation Figure 2 — InteractionC Timeline Visualisation

Conclusion

Despite Snapchat's design philosophy of disappearing content, this study confirmed that significant forensic evidence persists — the question is whether investigators have the access level and tooling to reach it.

Timeline Reconstructed
By cross-correlating three independent artefact sources — network PCAP timestamps, SQLite chat records from the rooted Android, and iOS Keychain/InteractionC data — a complete account-to-account communication timeline was reconstructed across the January 22 – February 19, 2026 study window. No single source was sufficient alone.

Key Outcomes

  • 1Snapchat's QUIC + TLS 1.3 network stack effectively defeats MITM content interception; network analysis provides metadata, not message content.
  • 2Android access level is decisive: FFS extraction from a rooted device yielded near-complete Snapchat artefact recovery; logical extraction on a non-rooted device yielded nothing actionable.
  • 3iOS encrypted backups provide valuable metadata — account timelines, interaction records, location data — but do not expose chat content without a full file system acquisition.
  • 4Ephemeral data persists. The gap between what Snapchat's design intends to erase and what forensic tools can recover is substantial, and is gated primarily by acquisition method.