CheckReality Inc. Raised $2.6M to verify trust in the AI era.

PLATFORM

Six detection models. One API call.

CheckReality combines GAN detection, metadata forensics, pixel analysis, noise patterns, content credentials, and reverse lookup into a single endpoint. No single model catches everything, so we run them all.

How it works

Multiple layers, combined signal

Every image goes through six independent detection models in parallel. Each model specializes in a different type of manipulation. We combine their signals using ensemble scoring to produce a single confidence verdict that's more accurate than any individual model.

Architecture

Built for production

Sub-second response times via REST API. Process single images synchronously or submit batches of up to 1,000 images with webhook notifications. SOC 2 Type II compliant infrastructure with end-to-end encryption and zero data retention.

Integration

One endpoint, any workflow

Send an image, get a forensic report. Works with claims intake, loan origination, KYC onboarding, content moderation, legal discovery, and any workflow where image authenticity matters. SDKs for Python, Node.js, and Go.

Detection models

What we analyze

01

GAN Detection

Identifies images from DALL-E, Midjourney, Stable Diffusion, Flux, and 8+ other generators.

02

Metadata Forensics

Analyzes EXIF data, compression signatures, and editing history to detect post-processing.

03

Pixel Analysis

Examines noise patterns, edge consistency, and color distribution at the sub-pixel level.

04

Deepfake Detection

Multi-expert face analysis detects face swaps, reenactments, and blending artifacts.

05

Content Credentials

Verifies C2PA provenance data and coalition for content authenticity signatures.

06

Reverse Lookup

Cross-references against known AI-generated content databases and source repositories.