fatcousin
// case type

deepfake investigation (video / audio / image)

face-swap, voice-clone, identity-impersonation. PRNU + GAN fingerprint + ELA + lip-sync + audio splice.

tools
14
priority
H
processing
local · in browser

start here · primary tools

ordered. work top-down. the first tool is the suggested entry point for this case type.

  1. video deepfake analyzerdrop a short video · sample frames · blink rate · face boundary flicker · temporal inconsistency score · runs locally
  2. face swap artifact detectordrop an image · jawline color mismatch · compression boundary heuristics · face-region signal table · runs locally
  3. gan fingerprint detectordrop an image · fft grid periodicity · color histogram anomalies · synthetic likelihood score 0–100 · runs locally
  4. ai synthetic voice generation artifact analyzeranalyze synthetic voice generation artifacts and identify possible AI-generated speech characteristics · spectrogram consistency, prosody, splice boundaries · runs locally
  5. audio splice detectordrop audio · spectral flux discontinuities · candidate splice timestamps · timeline table · runs locally
  6. ela image tampering detectordrop a JPEG · error level analysis · detect localized re-compression · flag tampered regions · visualize ELA map · runs locally
  7. prnu fingerprinterjpeg png sensor noise · residual fingerprint · pearson correlation · heatmap viz · LIKELY DIFFERENT · CSV · runs locally
  8. copy-move forgery detectordrop an image · block-matching copy-move scan · suspicious region overlay · heuristic clone map · runs locally

also useful · secondary tools

supporting and follow-up tools. surface as the investigation widens.

  1. video frame timestamp analyzerdrop an MP4 or WebM · step through frames · measure actual frame intervals · detect irregular timing · cut/splice indicators · frame drops · export timeline · runs locally
  2. video container forensic analyzerdrop mp4 mov mkv avi files · parse container atom box structure · extract all metadata tracks · chapter marks · embedded thumbnails · creation tool information · detect container manipulation · timestamp inconsistencies · runs locally
  3. audio edit detectordrop a WAV file · analyze spectral discontinuities · RMS energy shifts · noise floor changes · DC offset jumps · flag suspected splice points · runs locally
  4. metadata consistency checkerdrop a JPEG · PNG · or video · compare EXIF date vs file date vs filename date vs GPS timestamp · flag inconsistencies · dimension mismatches · runs locally
  5. electric network frequency audio authenticatordrop wav/mp3 · extract enf drift · splice checks · starter reference correlation · investigative only · runs locally
  6. case report generatorfill in case number · examiner · dates · findings · drop evidence files for auto hash · generates structured forensic report PDF · runs locally
// case-kit pipeline

run as a stack

skip the click-through. these presets are curated forensic pipelines you can save as a stack with one click and run on your evidence locally.

  • deepfake — image evidence kit

    7 steps

    drop suspect images → AI provenance → face-swap artifacts → copy-move forgery → LSB/chi-square stego → report

    1. 01evidence-manifest-generatorhash every image file before pixel analysis
    2. 02ai-generated-image-provenance-analyzermodel / generator fingerprint hints from metadata and artifact patterns
    3. 03gan-fingerprint-detectorFFT grid periodicity + histogram anomalies — synthetic image screener
    4. 04face-swap-artifact-detectorboundary and blending artifacts around swapped facial regions
    5. 05copy-move-forgery-detectorcloned-region detection — common in composite fakes
    6. 06image-stego-detectorchi-square / RS / LSB analysis on decoded pixels
    7. 07case-report-generatordraft a report suitable for legal review of still-image authenticity

  • deepfake — video temporal kit

    3 steps

    drop short MP4/WebM clips → sampled frame temporal metrics → blink rate + boundary flicker → report

    1. 01evidence-manifest-generatorhash every video clip before frame sampling
    2. 02video-deepfake-analyzersample frames locally — blink rate, boundary flicker, temporal score
    3. 03case-report-generatordraft a report on temporal face metrics for the clip set

  • deepfake — audio authenticity kit

    5 steps

    drop suspect audio → ENF grid analysis → splice detection → LSB stego screen → report

    1. 01evidence-manifest-generatorhash audio files before spectral analysis
    2. 02enf-audio-authenticatormains hum (ENF) presence + splice markers — investigative screening only
    3. 03audio-splice-detectorspectral-flux discontinuities suggesting edit points
    4. 04audio-stego-detectorLSB / chi-square / echo analysis on decoded PCM
    5. 05case-report-generatordraft a report summarizing ENF + splice + stego signals

  • deepfake — camera PRNU kit

    3 steps

    drop 2+ JPEG/PNG from claimed same camera → sensor noise fingerprint → pearson correlation → report

    1. 01evidence-manifest-generatorhash every image before PRNU analysis
    2. 02prnu-fingerprinteraccumulated noise residual K + per-image correlation — probabilistic screener only
    3. 03case-report-generatordraft a report on whether images likely share the same sensor
// pattern-matched

related tools

tools that the manifest-classifier flagged as plausibly useful here but that aren't in the hand-curated lists above. less editorial weight — scan, don't work top-down.

ready