Skip to content

Choose the model. Keep the work.

A local-first AI harness for chats, code and projects. Work in the browser, use cloud models, or connect this PC.

Try a demo of Locale

Gemini Nano isn't available in this browser.

DemoFree

Choose the model

Switch models, ask for consultation, or build a group.

In Chrome
Try Gemini Nano on supported devices.
This PC
Connect local models, folders and approved tools.
Locale Inference
Use an account balance for cloud models.
Connections
Bring provider access and API keys.

Example models for local use: Qwen3 4B, Llama 3.2 3B, Gemma 3 4B and Qwen3 8B, through the free Agent. Set up the Agent from inside Locale, when you want local models and folders.

Model rows from the app

One project, many chats

Keep instructions, summaries and files beside the work.

Review the changes
Review tool activity and control access to device folders.
Example project overview

Get started

Request an invite

  1. Try Locale

    The first time, setup picks a model and opens a chat with it. Chats, projects and browser files are free in this browser.

  2. Add this PC

    The free Agent for Windows and Linux runs local models and opens your folders. Its page in Settings has the download.

  3. Keep your work

    Chats stay in this browser for free. Keep, Sync and Cloud save chats and files to your account and sync them across your devices.

Request an invitation to use Locale with your own chats, projects and Agent. The demo above is available without an account.

Inference includes sponsored placements. Credit packs are also available without a subscription.