Skip to main content
v2026.11,858 entries · CC-BY 4.0
LAC HealthMedical & Research SuppliesReagents, PPE & instruments — chain-of-custody documented.Fast, traceable sourcing built for regulated research environments, from bench consumables to instrumentation.Shop lac.us CodeCASRAIlac.us

Forensically: Free Browser Image Forensics Tool Explained

What the Forensically image-forensics tool does, panel by panel, how to use it on a figure, and its stated limits.

Written and maintained by CASRAI Editorial Board

Last updated

Forensically is a free set of image-forensics tools that runs in a web browser. It was written by Jonas Wagner and is hosted at 29a.ch. Research-integrity staff use tools of this kind to take a closer look at a figure, but the tool’s own page says it works like a magnifying glass and cannot by itself show that an image is authentic or manipulated.

What Forensically is

The tool page describes Forensically as free, browser-based digital image forensics tools. It bundles several analysis panels in one interface: you open an image and switch between panels to inspect it. The word itself is also the English adverb of forensic, meaning by forensic methods or in a forensic manner. This page is about the tool, not the word.

In research integrity, the usual use is screening a published or submitted figure (a western blot, micrograph or photograph) for signs such as duplicated regions. For the wider workflow, see detecting image manipulation in figures.

Who makes it and what it covers

The official page at 29a.ch/photo-forensics credits the tool to Jonas Wagner, who also published a series of blog posts explaining individual panels (clone detection, noise analysis, principal component analysis, and JPEG forensics). The page does not state a licence or link to source code, so this guide does not describe the tool as open source.

The panels listed on the official page are in the table below.

Panel What the official page says it does
Magnifier Shows small hidden details.
Clone Detection Highlights similar regions within an image; similar regions are marked in blue and linked by red lines.
Error Level Analysis Compares the original image with a recompressed version; manipulated areas may look darker or brighter.
Noise Analysis Uses a separable median filter to isolate noise; works best on high-quality images.
Level Sweep Sweep the histogram with the mouse wheel to find discontinuities at edges.
Luminance Gradient Shows how brightness changes along the x and y directions.
PCA Principal component analysis; slow on large images.
Meta Data, Geo Tags, Thumbnail Analysis Reads EXIF data, GPS data if stored, and any hidden preview image.
C2PA Content Authenticity Reads signed content-credential metadata, which the page says is not inherently trustworthy; remote fetching is disabled for privacy.
JPEG Analysis, String Extraction Shows JPEG comments, quantization tables and marker structure; extracts readable ASCII strings, like the Unix strings command.

How to use it

  1. Open the official tool page at 29a.ch/photo-forensics in a current browser such as Firefox or Chrome.
  2. Use the Open File button in the header to load an image from your computer. The Help button sits next to it.
  3. Pick a tool panel. For a figure screened for reuse, Clone Detection is the usual starting point.
  4. In Clone Detection, matches that fall below the Minimal Cluster Size setting are discarded as noise, according to the author’s blog post. The post also describes a Show Quantized Image option that displays the compressed blocks the matcher works on.
  5. For Error Level Analysis, the author advises matching the JPEG quality setting to the original image’s quality.
  6. Compare what each panel shows with the original image, and treat every highlight as a lead to check, not a finding.

The author’s blog post says clone detection was a first attempt and not very refined, and that the built-in help is weak. A tutorial video is linked from the official page.

Strengths and limits

Strengths

  • Several complementary views in one tool, with no account.
  • Privacy: the FAQ says images are never uploaded and all images stay on your computer. It can also work offline in a modern browser.
  • Free of charge.

Limits

  • It cannot decide authenticity. The official page warns that results can be misleading. Bellingcat’s toolkit entry for the tool gives similar advice: do not rely on Forensically alone to decide whether a photo is manipulated.
  • RAW files are not supported; the page names 24-bit PNG as the highest-quality format it accepts.
  • Some panels are slow on large images (PCA) or work best on high-quality images (Noise Analysis).
  • Findings need an expert reading and a documented process. See responsible use of image detectors in research-integrity investigations.

Alternatives

Option What the sources say Where to read more
Ghiro Recommended on the Forensically page as other software; open source, with an online version at imageforensic.org. 29a.ch/photo-forensics
FotoForensics The Forensically page links its error level analysis tutorial. fotoforensics.com/tutorial-ela.php
Proofig Dedicated scientific image-integrity software, covered in a CASRAI guide. Proofig guide
ImageTwin Dedicated scientific image-integrity software, covered in a CASRAI guide. ImageTwin guide

FAQ

What does the word “forensically” mean?

It is the adverb form of “forensic”: by forensic methods, or in a way suited to forensic examination, for example an image examined forensically. As a proper name, Forensically is the browser tool described above.

Does Forensically upload my images?

According to its FAQ, no. Images are not uploaded and stay on your computer.

Can Forensically prove an image was manipulated?

No. The official page says it works like a magnifying glass and that results can be misleading. Treat output as a prompt for further review.

Which file formats work best?

RAW is not supported. The page names 24-bit PNG as the highest-quality format; error level analysis relates to JPEG compression.

What tools do forensic scientists use?

This page does not survey forensic science as a whole, which spans many disciplines with different instruments. For digital images specifically, the tools covered here are the ones named in its sources: the Forensically panels in the table above (clone detection, error level analysis, noise analysis, metadata and JPEG structure), the alternatives Ghiro and FotoForensics, and dedicated research-integrity software such as Proofig and ImageTwin.

What is forensic speaking?

The page’s sources do not define “forensic speaking”, and it is not a term this page uses, so no definition is given here. It is not a feature of the Forensically tool, which analyses images only.

What are the 7 principles of forensic science?

There is no single authoritative list of seven principles in the sources used for this page, so none is asserted here. The practical principles this page does rest on are stated in the tool’s own documentation: results can be misleading, and a highlight is a lead to check against the original image, not a finding.

Sources

  • Jonas Wagner, Forensically tool page, 29a.ch/photo-forensics (read 2026-11-07).
  • Jonas Wagner, “Forensically, Photo Forensics for the Web,” 29a.ch, 16 August 2015 (read 2026-11-07).
  • Bellingcat Online Investigation Toolkit, entry for Forensically (bellingcat.gitbook.io/toolkit/more/all-tools/forensically).

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about Forensically: Free Browser Image Forensics Tool Explained

Ask your first 2 questions free below. Subscribers get 150 a day for $29 a month.

An AI assistant specialized in research administration. It cites the sources behind every answer, labels web answers and says when it can't answer.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays free to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →