What Virtues is

Virtues is a server that holds the data of your life — messages, calendar, health, location, transactions, recordings, files — on a machine you own, and the software that turns that pile into something readable: a record of your days, articles about the people and places in them, and an AI that reasons from your actual life rather than a generic profile.

It runs on your own Linux machine today. Purpose-built hardware comes later; nothing on this site requires it.

What runs where

This is the part most people want answered first, so here it is plainly.

Your records stay on the server. Sources are pulled in and written to a database and a file store on your own disk. Nothing syncs them out. The embedding and reranking models that make your record searchable also run on the server, locally, on ports nothing outside it can reach.

The language models that write do not run on the server. When Virtues composes an account of your day or answers a question, the material for that request goes to a model provider — through our gateway by default, or through any OpenAI-compatible endpoint you point it at, including a provider you have your own account with. We meter what a request cost and never keep what was in it.

No inbound port is opened at home, and we have no way in. Your devices reach the server by key, on a list the server itself keeps. We aren’t on that list. Reaching your server describes the paths and states exactly what our relay can and cannot see, without rounding it up.

Where to start

If you’re installing on your own hardware, three pages carry the setup, in order. What to run it on says what the machine needs and, more usefully, what decides whether it feels fast. Setting up inference covers the one part that is genuinely yours to run — the embedding and reranking models your record is searched with — and is worth doing before you install, because the installer asks for their URLs first. Then Installing is one command and the ten minutes around it. After that, Reaching your server covers pairing a phone or laptop, which is how you actually use the thing.

Once it’s running, four pages carry the weight: Upgrading, Backup & restore — read that one before you need it, since the server deliberately cannot decrypt its own archives — When something breaks, and The CLI.

The Glossary defines the words this system uses for the parts of your life it holds, which is worth ten minutes if the vocabulary feels invented. It partly is.

What’s here, and what isn’t

These docs are written alongside the software and describe what actually ships. Pages marked soon in the sidebar are planned; they land as the features they cover settle, rather than in advance of them. Nothing published here is aspirational — if a page says the server does something, it does.

Every page here is also plain markdown: append .md to any docs URL to get the source, and /llms.txt indexes the lot for anything reading on your behalf. The manual is versioned with the software it describes — this site publishes from the released branch, so what you read is what a server actually runs.

The engineering record behind all of this — design decisions, audits, measured findings — lives in the repository rather than here. It is written for the people and agents building Virtues, and it reads that way.

The Library holds the essays — what we think this is for. That’s the other register entirely.