AI Model Inspector reads GGUF locally

This is a narrow but genuinely useful local-AI utility: less flashy than a model runner, but directly aimed at the annoying metadata-checking chores around downloaded weights.

• –Browser-only parsing is the right privacy posture for huge local models and proprietary ONNX files.
• –GGUF v2/v3 plus ONNX support covers both LocalLLaMA-style quantized models and broader deployment artifacts.
• –JSON export makes the proposed model-compare feature less urgent, though built-in diffing could still help users who do not want to juggle files in an editor.
• –The main risk is browser memory and malformed-file handling, which the project already flags as a limitation.