An AI tool that identifies whether an audio, video, or image is real or synthetic.
Drop an image, or browse
JPG · PNG · WebP · BMP · TIFF · HEIC · up to 4.5 MB
Create a free account for 50 checks a month and files up to 12 MB.
Upload a video, image or audio file. The detector scores it for signs of AI generation or manipulation and returns a verdict with a confidence score you can weigh alongside other evidence.
Detect deepfakes in audio, video, and image formats
The same detection through a REST API, from the Starter plan
An AI score with voice segments, video frame scores and an image source check
Names the AI generator when an image still carries C2PA or metadata
No training, no tuning, no integration required. Upload a file and read the verdict. For the API, see the docs below.
Free tier ships with 50 detections / month, image, audio, and video detection.
Drag and drop a video, image or audio file into the web app.
A verdict and an AI score from 0 to 100, with voice segments, video frame scores and an image source check. Export your history as CSV on paid plans.
Whether you're protecting a household from voice-clone scams or a finance team from a $25M wire fraud, the detection problem is the same. The stakes aren't.
People have lost money to deepfake voice scams after receiving messages that cloned the voices of their loved ones.
With our Deepfake Detector, you can check the voice messages and call recordings you receive. It gives you a fast way to check them, with an AI score to guide what you do next.
Protect Yourself Now →Businesses lose large sums to scammers who clone the voice of a senior executive to authorize fraudulent transfers. It has happened before, and these attacks are increasing.
Video, image, and voice each leave different fingerprints. We run a purpose-built detector for each, all in one account.
Scores sampled frames for signs of AI generation and keeps the highest frame with the result. Clips up to 2 minutes free and 10 minutes on paid plans.
Detect video →Identify AI images from Midjourney, DALL·E, Stable Diffusion, Flux, Firefly and other generators, with a verdict, an AI score and a source check that names the generator when the file still carries its metadata.
Detect images →Detect synthetic speech from ElevenLabs, Resemble, PlayHT, OpenAI, and other voice-cloning engines, scored in 6-second segments so you see where the voice turns synthetic.
Detect voice →Deepfake detection identifies audio, video and images that were generated or altered by AI. Whether it's a cloned voice or an AI-generated photo, the detector scores the file and tells you how likely it is to be synthetic. Learn how deepfake detection works.
Dedicated detectors: AI Video Detector · AI Image Detector · AI Voice Detector
Test Your Files →Skip the UI. Wire detection directly into your moderation pipeline, your KYC flow, or your fraud team's case manager. Built for engineers who'd rather curl than click.
Get a signed upload URL, send the bytes straight to storage, then one call returns the verdict and AI score.
Image results include the source check; voice results include 6-second segments. For video at volume, talk to sales.
Bearer-key auth on Starter and above. Keep your sk_ key on your server, never in client code.
Full access to our Deepfake Detector. Protect yourself from audio, video, and image manipulation. Cancel anytime.
Three documented cases. Three different attack surfaces. Each one started with a voice or a face the victim trusted.
A woman in Hyderabad lost Rs 1.4 lakh to a scammer using AI to mimic her nephew's voice, claiming he needed money due to an accident. The scam, involving a late-night call and a plea for secrecy, deceived her into transferring funds.
A finance worker at a multinational firm was tricked into sending $25.6 million after a deepfake video call with fraudsters posing as the CFO and colleagues. The scam was uncovered after verification with the company's head office.
Jennifer DeStefano received a call claiming her daughter Brianna was kidnapped, demanding a $1 million ransom. The call turned out to be a scam using a cloned voice. Brianna was safe the whole time.
One account checks video, images and voice, and every result shows its working: an AI score, voice segments, video frame scores and an image source check.
Our detector helps estimate whether audio, video, or an image was AI-generated, across voice cloning, video synthesis, and image generation models, not limited to a specific platform.
Voice segments, video frame scores and the image source check show where a result comes from.
Image and voice detection over a REST API with JSON responses, from the Starter plan.
The questions we get asked most about detecting AI video, images, and voice. Want the deep dives? The blog goes further.
Video, images and voice, all in one account. Upload a clip, photo or audio file and get an AI score from 0 to 100 with a verdict: Human, Likely Human, Likely AI or AI Generated. Voice is scored in 6-second segments, video frame by frame, and images get a source check.
Detection is probabilistic, so every result is an AI score and a verdict, not a guarantee. Results are strongest on original files; compression, screenshots and screen recordings make any check harder. Weigh the score alongside the source, the context and the original upload.
Yes. A free account includes 50 checks a month across video, image and voice, with no card required. You can also check images with the free checker on this page, 10 a day with no signup.
Images: JPG, PNG, WebP, BMP, TIFF and HEIC. Audio: MP3, WAV, OGG and M4A. Video: MP4, MOV and WebM. Free accounts check voice and video clips up to 2 minutes; paid plans go up to 10 minutes.
It is built for current generators: video from tools like Sora, Runway, Pika and Kling, images from Midjourney, DALL-E, Stable Diffusion, Flux and others, and voices from ElevenLabs, PlayHT, Resemble and OpenAI. Every result is a probability, not a guarantee.
Trained models score the file for signs of AI generation and return a probability, not a guess. Images also get a source check that reads C2PA and other metadata. Our deepfake detection guide walks through how it works.
Field guides on how deepfakes are made, how to detect them, and how to stay ahead of synthetic-media scams.
Education
Education
ScamsStart free. Upgrade to Starter, Business, or Enterprise when you need more. Secure payment with Stripe. Cancel anytime.