Detect Midjourney AI Images
Upload an image and find out whether Midjourney made it. FauxLens covers v4, v5, v6 and Niji outputs, shows you what it found, and answers in seconds. Free to try, no sign-up.
Upload the image or video. It is deleted after the check.Drop a photo or video here, or click to upload. It is deleted after the check.
JPG PNG HEIC · MP4
Detects content from
- OPENAI
- MIDJOURNEY
- STABILITY
- GEMINI
- FIREFLY
- FLUX.1
- CANVA
- IDEOGRAM
How every scan runs
Read the methodProvenance
C2PA content credentials and EXIF metadata, checked first
Watermark & Pixels
OpenAI and Google AI watermarks read, then a detection model scores the image
Evidence Chain
Every finding itemised and graded, not reduced to a yes or no
- Files deleted after analysis
- No account required
How to detect a Midjourney image
Midjourney is the hardest of the major generators to catch by eye, which is exactly why eyeballing it stopped working. FauxLens checks a suspected Midjourney image in stages. First it reads the file itself: C2PA content credentials and EXIF or XMP metadata. If a generator signature survives there, the scan can stop on it, and if the metadata is gone, that absence is never counted against the image. Second, it looks for the invisible watermark OpenAI and Google embed in images their own tools make, which would point to one of those vendors instead. Third, a specialist AI-image detection model scores the pixels and judges whether the image is real, partly changed by AI or generated by AI, with a multimodal vision model as the fallback. The pixel model does not name Midjourney; it tells you whether AI made the picture. Results come back as a verdict with a confidence score and the evidence lines behind it, so you can see what drove the call.
What makes Midjourney output distinctive, version by version
Midjourney uses a proprietary cascaded diffusion pipeline - separate networks handle low-frequency structure and high-frequency detail in sequence. This architecture produces a distinctive artifact pattern that differs meaningfully from Stable Diffusion's single latent-space approach or DALL-E's CLIP-guided generation. The tells you can check yourself are: systematic over-smoothness in skin, fabric, and foliage textures caused by the model's learned bias toward a polished look over photographic accuracy; a glossy, evenly lit finish that real cameras rarely produce; and lighting that is convincing at a glance but wrong on inspection, with shadow directions that contradict the light source and eye reflections that do not match the scene. These are visual clues, not proof. FauxLens relies on metadata, watermarks and a detection model that scores the pixels, which still works on files that have been compressed or resized moderately.
Read moreShow less
Each Midjourney major version introduced architectural changes that shifted its forensic profile. Version 4 (late 2022) used an earlier diffusion backbone that produced more visible checkerboard artifacts in uniform color regions - detection accuracy on v4 images approaches 99% because these artifacts are strong and consistent. Version 5 (early 2023) represented a substantial quality leap: the upscaler was redesigned, photorealistic portraiture improved dramatically, and the telltale artifacts became subtler. Detection accuracy on v5 outputs is approximately 96%. Version 6 (late 2023 to present) added coherent text rendering, a capability Midjourney previously lacked, and introduced a new attention mechanism that gives text and fine-detail regions more for a detector to catch, making v6 outputs containing text somewhat easier to detect than pure portrait outputs. Niji mode, Midjourney's anime-specialized variant developed in collaboration with Spellbrush, uses a fine-tuned model that shares the base v5/v6 backbone but produces characteristic flat-shading and line-weight patterns. Niji images look quite different from standard Midjourney outputs, but the detection model judges whether AI made an image rather than its style, so anime-style Niji output is checked the same way.
Where Midjourney images turn up
Midjourney has become a primary tool for AI-generated profile photos used in romance scams, owing to its accessibility through Discord and the photorealistic quality of its portrait outputs. Criminal operations use it to generate consistent visual personas (typically an attractive professional in their 30s to 50s, sometimes a military officer or engineer) with multiple photos showing the same synthetic person across different settings. Unlike stolen real photos, these images return no results in reverse image searches, making forensic pixel analysis the only available detection method. In political disinformation, Midjourney images appeared in documented influence operations targeting the 2024 US election cycle and the 2025 European Parliament elections, used to fabricate crowd photos, protests, and situational imagery. Researchers have found Midjourney-generated content in influence operation image sets. On the commercial fraud side, Midjourney images have been used to create fake product listings, fabricated testimonial photos, and synthetic influencer personas monetized through affiliate marketing. AI-generated imagery now turns up routinely in fraud reports.
Getting a reliable answer
Upload quality significantly affects detection confidence. Original PNG or JPEG files exported directly from Midjourney's Discord bot or web interface keep their metadata and full pixel detail, and yield the highest confidence scores. Screenshots strip the metadata and add a compression round-trip that loses fine detail, reducing overall confidence. Images that have passed through social media platforms (Instagram, Twitter/X, Facebook) are aggressively re-compressed and resized, and heavily compressed or small copies are where the detection model is weakest. If you are analyzing a social media image, download the highest-resolution version available rather than screenshotting. For inconclusive results (confidence between 40% and 65%), try uploading a different version of the same image if available, the original file typically carries stronger signals than a downloaded social copy. Images processed through AI photo enhancers (Topaz, Remini, Let's Enhance) may show reduced confidence because these tools repaint fine detail and can blur the traces the detection model looks for. Images upscaled by Midjourney's own upscaler retain strong detection signals because the upscaler keeps the generated look of the original while adding detail.
Is that photo or video real?
Frequently asked questions
FauxLens detects Midjourney images across all versions with 95%+ accuracy. It reads the metadata, checks for provenance watermarks, and has a specialist detection model score the pixels, which still works after standard post-processing.
Detection accuracy exceeds 95% for v5 and v6 outputs and approaches 99% for v4. Portrait images from the latest versions are the most challenging category. Accuracy drops on small or heavily re-compressed copies, so the original file gives the most reliable answer.
Midjourney does not embed visible watermarks in generated images. Pro subscribers can generate without any visible branding. FauxLens reads any generator signature left in the metadata and has a detection model score the pixels, though heavy editing, compression or AI photo editors can weaken what it finds.
Yes, Midjourney v6 introduced a new attention mechanism and improved text rendering, and its output looks different from v5. V6 images containing coherent text tend to be easier to catch than portraits. Portrait-only v6 outputs are slightly harder to detect than equivalent v5 outputs. FauxLens does not identify the version; it tells you whether AI made the image.
Running a Midjourney image through a secondary AI tool, such as img2img in Stable Diffusion or another generator, introduces a second set of artifacts that can partially mask the original Midjourney signature. This is a known adversarial technique. FauxLens judges whether AI was involved, whichever generator was used last, though each extra pass can make the call less certain.
Niji mode is Midjourney's anime-specialized model, developed in collaboration with Spellbrush. It uses a fine-tuned version of the base Midjourney model optimized for flat-shading, strong line weights, and anime aesthetics. Niji images look very different from standard Midjourney output, but the detection model judges whether AI made an image rather than what style it is in, so Niji output from v4 through v6 is checked the same way.
Midjourney does not embed user attribution metadata in generated images. The Midjourney website indexes publicly generated images and allows searching by prompt - if the image was generated on a public server (not Stealth mode), you may find it by searching midjourney.com/explore. FauxLens can tell you whether an image is AI-generated, and reports the tool when the metadata names it, but cannot identify the specific account that created it.
Yes. Midjourney's own upscaler (both the standard and subtle options) keeps the generated look of the original while adding higher-resolution detail. Images upscaled with third-party tools like Topaz Gigapixel can show lower confidence, because those tools repaint the fine detail that the detection model looks at.
Keep exploring
Guides
- Can You Tell the Difference? Midjourney vs DALL-E vs Real Photos [Forensic Test]A technical showdown between the major AI engines. This guide breaks down the specific 'digital fingerprints' associated with Midjourney v6, DALL-E 3, and Stable Diffusion.
- Every AI Image Has a Hidden Fingerprint - Here's How Forensics Finds ItEvery AI model leaves an invisible fingerprint in the images it generates. GAN fingerprints are the forensic equivalent of ballistic markings-unique to the weapon, invisible to the eye, detectable by science.
- The Science of Deception: How AI Detection WorksA detailed guide to the digital forensics of AI image detection, from Error Level Analysis (ELA) and JPEG compression artifacts to Photo Response Non-Uniformity (PRNU), and which checks a FauxLens scan actually runs.
More tools
Detect while browsing: try the Chrome extension
Right-click any image · No account required