Tools

VRAM & context calculator

Weights + KV cache + overhead for any model, quant and context. Shows which devices it fits.

“Will it run?” hardware checker

Pick your GPU or Mac, a model and quant. Get a fit verdict and estimated tokens/sec.

Command builder

Generate llama.cpp, Ollama, vLLM and MLX commands with every flag explained.

Quantization explorer

See quantization error live and compare GGUF quants by size and quality.

Inference visualizer

Animated prefill/decode, speculative decoding and batching simulators.

In-browser playground

Chat with a small model running on your own GPU via WebGPU. Nothing is sent to a server.