FauxLens vs Pella Research

FauxLens analyzes images and video for signs of AI generation or manipulation, returning a verdict, confidence score, and forensic evidence. Its six-signal engine uses ELA, PRNU, GAN fingerprints, FFT, shadow physics, and EXIF forensics. Listed detection coverage includes OpenAI, Midjourney, Stability, Gemini, Firefly, Flux.1, Canva AI, Ideogram, Bing, and Grok. Image cleanup tools remove unwanted people, visual noise, and platform overlays. The web tool accepts JPG, PNG, and WebP images, and cleanup exports retain original dimensions up to 4K. A Chrome extension can scan publicly accessible images via right-click, but cannot scan images requiring authentication or blocked by cross-origin policy. The REST API accepts image or video files, or a URL, and returns structured JSON with verdict, confidence, and evidence by layer. API keys are limited to 20 requests per minute per key and 30 per minute per account; file requests are capped at 25MB and URL fetches at 50MB. FauxLens says uploaded media is deleted after analysis and scan history remains in the browser. Free includes 40 credits, described as four detections; one-time plans start at $2.99. API verdicts are informational, not proof or a sole basis for consequential decisions.

Pella Research studies synthetic speech and builds models for detecting it in audio. Its research examines vocoder artifacts, spectral statistics, and prosodic cues. A browser demo accepts an uploaded or recorded voice clip and returns a verdict without signup; the site says audio is not stored. The REST detection API returns calibrated authenticity scores and per-segment analysis, with access reviewed after a request, typically within 24 hours. Pella offers the API to journalists, platforms, researchers, and institutions, and lists a free tier for accredited journalists and academics. Published evaluation sets track new voice-cloning systems with documented provenance and no train/test leakage. Pella says models are trained and tested with phone-line compression, background noise, and re-encoding. It reports pellav2 results of 95.82% accuracy, 0.959 F1, and 57 ms latency in the August 2026 Podonos benchmark, plus 93.6% accuracy on m4a files with errors leaning toward false positives. pellav2 is described as open weights under the MIT license.