Welcome to Aivyx PA
Aivyx PA is a personal AI assistant that runs on your own hardware. It learns how you work, keeps its memory in an encrypted file on your machine, and records everything it does in a tamper-evident log you can inspect at any time. There is no Aivyx PA cloud — your assistant talks directly to whichever AI model you choose (a local model with no API key, or a provider like Anthropic or OpenAI).
This guide is the end-user manual. It walks you from first launch through every part of the Studio — the web interface you’re reading this in.
What makes Aivyx PA different
- It’s yours. Your data, your model choice, your machine. Nothing is sent to a hosted service you don’t control.
- It learns. Beyond the personality you give it at setup, your assistant refines how it helps you over time — and you approve every change.
- It’s accountable. Every action it takes is written to a cryptographically chained audit log. You can always see exactly what happened.
- It asks before doing anything risky. Sending an email, deleting a file, or spending money always stops for your confirmation.
How to read this guide
If you’re brand new, start with Getting started and then Create your agent — those two get you to a working assistant. After that, browse the pages for whichever features you want to use: the Studio screens, the terminal, autonomy and routines, security, and backups each have a page.
For exhaustive detail — every command, config setting and tool — see the reference manual.
The big picture
Aivyx PA runs as a small background service (the daemon) on your computer. Every way you talk to your assistant — this web Studio, the desktop app, the terminal, voice, or chat apps — is just a different window onto that same daemon. Close a window and the assistant keeps running; your conversation, memory, and settings persist.
You can also run several separate assistants side by side — say one for research and one for the household — each with its own memory, personality and Studio.