Local AI scanner
Find what can run on the machine in front of you.
Browser APIs can only estimate hardware, but they are enough for a useful first pass: cores, memory hints, GPU capability, and practical local model recommendations.
Best fit
Scanning this browser...
CPU cores
Scanning
Memory hint
Browser hidden
WebGL
Scanning
WebGPU
Scanning
Detected profile
- Platform
- Scanning
- Browser
- Scanning
- Device class
- Scanning
- GPU
- Not exposed by browser
Use this as a sizing guide, not a benchmark. Real performance depends on RAM pressure, thermal limits, quantization, GPU drivers, and the runtime you choose.
Recommended local models
Install with Ollama, LM Studio, llama.cpp, or a managed local runtime.
Llama 3.2 1B
Meta · 1.5 GB RAM target · 0.7 GB download
Fast local chat, summaries, simple writing tasks
Qwen 2.5 1.5B
Alibaba · 2.5 GB RAM target · 1 GB download
Light coding help, structured extraction, multilingual tasks
Phi-4 Mini
Microsoft · 3 GB RAM target · 2.3 GB download
Small reasoning, coding, agent utility work