See through synthetic reality.
TruthLens AI inspects every pixel — detecting AI-generated images, deepfakes, and AI edits with forensic confidence in under five seconds.
Drop an image to analyze
or browse files — JPG, PNG, WebP up to 6MB
Analysis Report
Verdict, confidence score and forensic breakdown will appear here.
Awaiting image…
Trusted by newsrooms, platforms & forensic teams
Capabilities
Built for the post-generative era
Six detection layers working together to surface the truth behind any image.
AI vs Real Classifier
State-of-the-art vision model distinguishes generated images from camera photos.
Deepfake Detection
Spots face-swaps, identity-replaced portraits, and AI-edited people.
GAN/Diffusion Artifacts
Detects melted hands, inconsistent reflections, weird text, and other tells.
Suspected Source Model
Estimates the likely generator: Midjourney, Flux, SDXL, DALL·E and more.
Forensic Reasoning
Returns human-readable evidence — not just a black-box score.
Sub-5 Second Latency
Streaming pipeline returns a verdict in seconds, not minutes.
Workflow
From upload to verdict in seconds
Three steps, zero setup. No API key required for the playground.
Upload
Drag, paste, or pick an image — JPG, PNG, WebP.
Analyze
Vision model + heuristic checks inspect pixels and patterns.
Report
Get a verdict, confidence, suspected model and reasoning.
98.4%
Detection accuracy
<5s
Average response
12M+
Images analyzed
150+
Countries served
Loved by professionals
What experts are saying
“TruthLens flagged a deepfake in our newsroom workflow in under 4 seconds. Game-changer.”
“We screen thousands of submissions daily. The API is fast, accurate and absurdly easy to integrate.”
“The reasoning output is what sold us — auditable evidence, not a magic number.”
Pricing
Simple, scalable plans
Pro
For creators & teams
$29/mo
- 10,000 scans / mo
- API access
- Reasoning + artifacts
- Scan history
FAQ
Frequently asked questions
How accurate is TruthLens AI?+
Our vision pipeline reaches ~98% detection accuracy on benchmark datasets covering Midjourney, SDXL, Flux, DALL·E and Gemini-generated imagery. Real-world accuracy varies with image quality.
Can it detect deepfakes?+
Yes — we specifically flag identity-replaced faces, AI inpainting on portraits, and inconsistent facial geometry that classic detectors miss.
Is my image stored?+
No. Images are processed in-memory for the analysis request and discarded immediately after. We never train on your uploads.
Do you offer an API?+
Pro and Enterprise plans include a REST API with the same response shape you see in the playground.