local guide

Install the Local Companion on macOS or Linux

Start and pair the device-local bridge without creating a site-owned model endpoint.

Cracked-gray Curious Obmil carries a glowing device toward a smaller, sorrowful GoldCap with one broad main cap, exactly two smaller caps, and living gold veins in a device-local workshop glade
Curious brings the work home. GoldCap keeps the training path on the user’s own ground.
PREREQUISITES
  • Python 3.11 or newer
  • A compatible local runtime
  • A downloaded local model
SUCCESS EVIDENCE
  • The endpoint remains on loopback
  • The browser authenticates with the token
  • The selected model answers a test prompt
ORDERED PROCEDURE

Complete the task

01

Download the current companion ZIP

02

Verify its SHA-256 hash

03

Extract it into a user-owned folder

04

Read README.txt

05

Run start-macos-linux.sh from that folder

06

Keep the token private

07

Pair from Workbench local settings

08

Run health and model-list tests

Warnings
  • macOS may require explicit permission to run the script
  • MLX training requires Apple silicon and a separate compatible installation