Model Attribution Pipeline
A custom RunPod-hosted classifier predicts the most likely base checkpoint and feeds downstream retrieval.
GenGlasses is an AI image attribution and retrieval system designed to investigate likely source models, cross-reference web evidence, and surface plausible model matches through a structured search workflow.
The application combines authentication, secure media handling, model inference, Google Lens evidence recovery, and Civitai retrieval into one focused technical tool.
A custom RunPod-hosted classifier predicts the most likely base checkpoint and feeds downstream retrieval.
Hive remains in the loop for AI-generated and deepfake scoring to provide an independent signal alongside attribution.
Google Lens evidence and LLM query planning are used to generate targeted Civitai searches and surface plausible matching models.