Validate integrity, compatibility and supply-chain risk for GGUF, Safetensors and ONNX — then compare two versions to classify and localize exactly what changed at the tensor level.
Scan what the file is. Compare what the model became.
See deliberately corrupted models — poisoned tensors, tampered weights, hidden backdoors — with real annotated scan reports in LLM Labs.
DEMO Scans without an account are public and visible here for 30 days, then automatically removed. Want a private report instead? Create a free account — you can delete any of your own reports anytime.