Runs on

llama.cpp

Local inference on the machine you carry. No hosted endpoint.

See features

Overview

PortableAI is the laptop-hosted line: you choose the local runtime and weights. llama.cpp-style local inference fits that model — GGUF files on disk, answers on the device.

What that means

  • Inference stays on the laptop.
  • No cloud inference bill and no vendor lock-in to a hosted API.
  • Pair the phone over local WiFi when you want a handheld client.

Not a SaaS integration

Nothing here phones home to generate a reply.