Skip to main content

Private, in-browser estimate

Can your phone run local AI?

Check your memory, storage, and workload before downloading a model. The result is a cautious starting point, not a performance promise.

Your starting point

Good starting point

This device has a practical starting budget for compact local models and ordinary chat.Leave storage headroom for the app, runtime files, indexes, and operating-system updates.

This estimate runs in your browser. Your selections are not uploaded or saved.

What the estimate can tell you

Local AI loads model data into memory and stores model artifacts on your device. More RAM and free storage give the runtime room to work, but they do not prove that a particular model, browser, or phone will be compatible. CuriousLM applies its own compatibility checks before enabling a download.

The safest approach is to choose the smallest compatible artifact that meets your task, leave storage headroom, and test a short prompt before importing documents or starting a long conversation.

Why the first answer may take longer

Before the first visible token, a local model may need to load, prepare the prompt, and process prior conversation text. Longer prompts and document context can increase that delay. Closing heavy apps and reducing unnecessary context can help on memory-constrained phones.

Heat and battery limits can also reduce performance over time. A quick test is useful, but sustained use is the better measure for your real workflow.

Continue learning

Choose with evidence, not a model-size slogan.