OFFLINE AIlocal · private
CLOUD CHAT

A LITTLE INDEPENDENCE

Your ideas. Your device.

Choose a local model, make a little room, and keep thinking wherever you are.

PERSONACREATIVITY 0.7

Models download once into this browser and run on this device's processor. Nothing you type leaves the device. On a phone, start with a model under 1 GB.

Qwen 3.5 0.8B · 0.58 GB
Smallest. Any phone.
LFM 2.5 1.2B · 0.73 GB
Fast everyday chat.
LFM 2.5 1.2B Thinking · 0.73 GB
Thinks before answering.
Llama 3.2 1B · 0.81 GB
Meta, quick and steady.
Gemma 3 1B · 0.81 GB
Google, small.
Qwen 3.5 2B · 1.4 GB
Smarter; newer phones.
MiniCPM 5 2B · 1.56 GB
Strong for its size.
LFM 2.5 2.6B · 1.67 GB
Best LFM; newer phones.
Llama 3.2 3B · 2.02 GB
Laptops and high-end phones.
Granite 4.1 3B · 2.1 GB
IBM. Laptops.
Gemma 4 E2B · 3.46 GB
Computers only.
wednesday
pick a model above

Private chat with Wednesday, running on your own hardware.