Model selection
Small models
Designed for lighter downloads and everyday tasks on more devices.
Local model marketplace
PocketBrain helps you explore compatible models by size and purpose, then manage them locally.
Model selection
Designed for lighter downloads and everyday tasks on more devices.
Model selection
A balanced option for richer responses and practical local workflows.
Model selection
For capable hardware and work that benefits from more model capacity.
Manage locally
Download models when you need them, pause or resume transfers, retry downloads, switch installed models, and remove models you no longer want.
Select deliberately
Task
Start with the kind of writing, chat, or file work you want to do.
Hardware
Memory, storage, and native runtime affect which models are practical.
Control
Install more than one model, switch when needed, and delete what you no longer use.
The PocketBrain approach
After a compatible model is installed, core workflows can continue without an internet connection.
PocketBrain has no cloud chat backend and does not require an account for core use.
Your models, local storage, and files remain under your control on the device.
Before you begin
PocketBrain works best when your device and model choice are planned together.
Common questions
PocketBrain’s capability is shaped by its local model, native runtime, and your device. Start here, then explore the complete FAQ.
Read the full FAQ →Core compatible workflows can work offline after a model is installed. What is available depends on the model, device hardware, and native runtime.
Keep exploring