Use cases
Not every AI feature belongs in the cloud. These are the shapes of app where running the model on the phone — with llamafu — is the better engineering and product decision.
Who
Apps handling personal notes, health, journaling, or finance.
When the prompt contains the user’s most sensitive data, the safest place to process it is the device it already lives on. With llamafu, text never touches a server — there is no API to breach and no log to subpoena.
Who
Aviation, maritime, remote inspection, disaster response, travel.
Cloud AI disappears the moment connectivity does. On-device inference keeps summarisation, translation, and Q&A working with the radio off — on a plane, underground, or in the middle of nowhere.
Who
Healthcare, legal, government, and enterprise with data-residency rules.
Data-residency and confidentiality requirements often forbid sending user content to third-party APIs. Keeping inference on the device sidesteps the transfer entirely and shrinks the compliance surface.
Who
Consumer apps where an AI feature ships to millions of users.
Per-token API pricing punishes success: the more people use the feature, the bigger the bill. On-device inference moves compute to the user’s hardware, so the marginal cost of each generation is zero.
Start with the quickstart, or tell us what you’re shipping.