Skip to main content

Memory-first local AI assistant

Build around memory you own

Use these docs to install the app that owns the memory, connect the browser side panel, configure providers, and keep one local Brain across terminal, tools, channels, cron, and subagents.

Install appPair browser
memory-firstlocal graph memory, not one model accountportableswap models without rebuilding contextdaily surfacesbrowser, launcher, terminal, cron, skills, and IM share one Brain

Documentation map

Start with the local app, then connect every surface to the same Brain.

Move from install and provider setup into browser pairing, terminal workflows, memory, subagents, channels, voice, and local data boundaries.

01

Install the app

Use the launcher, release installers, or CLI paths to set up the daemon, provider, model, and local home directory.

02

Pair the browser

Generate a gateway URL and bearer token, then connect the side panel to the same local runtime and Brain.

03

Understand Brain

Learn how Brain forms, recalls, and maintains a Cognitive Nexus for projects, preferences, relationships, and decisions.

One local runtime

Browser, launcher, terminal, files, cron, skills, and IM channels can feed one memory thread.

launcher

setup, status, pairing, logs, restart, and updates

browser

page context, screenshots, downloads, and approved actions

terminal

commands, files, skills, subagents, and local workspaces

channels

Telegram, WeChat, Discord, Lark/Feishu, and voice