AI Model Inspector
Reads the metadata inside a GGUF, Safetensors or ONNX model file — architecture, tensor shapes, producer info — without downloading the full model into a Python environment first. The format is detected automatically.
Drop a .gguf, .safetensors or .onnx file here, or click to choose one
Works entirely offline in your browser. Nothing is uploaded — and only the header/metadata is actually read, not the (often huge) tensor data itself.
When you'd use this
- Checking a downloaded model file's architecture and parameter count before loading it
- Reviewing a GGUF, Safetensors, or ONNX file's metadata without loading the full model
- Verifying a model file's quantization or format before deploying it
For ONNX files, this reads top-level model metadata and the graph’s input/output names only — it doesn’t walk the full list of operator nodes. GGUF and Safetensors show the complete tensor list (name, shape, dtype) alongside all metadata.