Does it load the whole model?

No. Only the header at the start of the file is read, usually a few MB, so a 70 GB model opens in about a second and nothing is uploaded.

Can it tell me if a model fits in my memory?

It shows the parameter count and the quantization. File size is a good guide to the memory a model needs, plus some room for the context.

Are my files uploaded?

No. This tool runs entirely in your browser tab. Files are read from your disk into memory and the result is generated locally.

Is it free? Are there limits or watermarks?

Free and open source (MIT). There are no file-size caps, daily quotas, accounts or watermarks.

Does it work offline?

Yes, once the page (and any engine it needs) has loaded, it keeps working without a connection thanks to the service worker cache.

Home / AI Builder / GGUF & Safetensors Inspector

GGUF & Safetensors Inspector

Read a GGUF or safetensors model header: architecture, quantization, context length, parameter count, tokenizer and every tensor, without loading the weights.

Free, no uploads, no sign-up. Runs 100% in your browser.

How GGUF & Safetensors Inspector works

Everything runs in your browser. Prompts, datasets and model files are never uploaded. Only the start of the file is read. GGUF headers (versions 1 to 3) are parsed key by key, including long arrays such as the vocabulary, which keep their size and first items. Safetensors files start with an 8 byte length and a JSON header listing every tensor. Even very large models open in about a second.

Input: GGUF, SAFETENSORS. Output: JSON, CSV.

How to use GGUF & Safetensors Inspector

Drop a .gguf or .safetensors file.
Read the summary: architecture, quantization, context length and parameters.
Filter the metadata and tensor tables.
Download the header as JSON or the tensor list as CSV.

Tips

Bits per weight tells you how strongly a model is quantized: about 4.5 for Q4_K_M, 8.5 for Q8_0, 16 for F16.

GGUF & Safetensors Inspector FAQ

Does it load the whole model?: No. Only the header at the start of the file is read, usually a few MB, so a 70 GB model opens in about a second and nothing is uploaded.
Can it tell me if a model fits in my memory?: It shows the parameter count and the quantization. File size is a good guide to the memory a model needs, plus some room for the context.
Are my files uploaded?: No. This tool runs entirely in your browser tab. Files are read from your disk into memory and the result is generated locally.
Is it free? Are there limits or watermarks?: Free and open source (MIT). There are no file-size caps, daily quotas, accounts or watermarks.
Does it work offline?: Yes, once the page (and any engine it needs) has loaded, it keeps working without a connection thanks to the service worker cache.

Related tools

Token Counter: Count tokens with the real GPT-4o, GPT-4, Llama 3, Mistral, Qwen, Gemma or DeepSeek tokenizer, see each token, context use and cost.
File Type Identifier: Find out what a file really is from its magic bytes, with entropy and header view.
Hex Viewer: View any file as hex and ASCII, search bytes, inspect values and extract strings.
MCP Config Builder: Build Model Context Protocol server configs for Claude Desktop, Cursor, VS Code, Windsurf, Claude Code and Gemini CLI, or convert one app's config to another.