virtualme / q & a text our agent
15 questions, plainly answered

questions, mostly the nervous ones.

we built virtualme to be quiet, ownable, and small. these are the questions people ask before they're willing to text a stranger their thoughts.

the basics

01

what is virtualme, in one line?

a personal wiki — written for you, with you — that every ai you use can read from. so you don't have to re-explain yourself in every chat.

02

do i need to install anything?

no. it's a phone number you text, and a small connector you paste once into chatgpt or claude. that's it.

03

why text? why not an app?

because you already text. because we don't want to be a tab open in your browser. because conversations are how memory actually grows — slowly, in fragments, when something occurs to you.

privacy

04

who can read my pages?

you, by default. nobody else, unless you flip a page public — and even then, only that page. there's no shared timeline, no feed, no "discover".

05

do you train ai on my notes?

no. never. your wiki is read by the agent you sign in with, scoped to your conversation. it isn't fine-tuning data, and it isn't pooled with anyone else's.

06

can i delete everything?

yes. one tap wipes the vault server-side. before you do, take an export — the markdown stays in your downloads forever.

your data

07

where is the markdown actually stored?

we keep a copy on our servers so the agents have something to read. you keep a copy whenever you tap export. the format is plain .md, frontmatter when the page has metadata — no proprietary container.

08

can i export and leave?

that's a feature, not a fight. one tap downloads a tarball (.tar.gz) of your vault. macOS and linux open it natively; windows users may need 7-zip. drag it into any markdown app and keep going.

09

does it work with obsidian?

yes. the export is obsidian-compatible — backlinks, frontmatter, tags, folders. logseq too.

the agents

10

does it work with chatgpt?

yes — paste the connector once into a new chatgpt project. from then on, your wiki is in context for every chat in that project.

11

claude? gemini? others?

claude works the same way (a project + connector). gemini, mistral, and a handful more are wired in. if your favorite isn't listed yet, ask — we add things slowly but steadily.

12

what happens the first time my ai reads the prompt?

it reads your wiki, then asks five short questions to seed what's missing — your name, your work, the people you care about, what you're chasing, anything else. answer in your own voice. then come back to your wiki and watch the pages appear.

13

can the ai write to my wiki without me?

no. never. every write is a diff that lands in your inbox. your tap is the gate. that's the whole shape of the product.

money

14

is it free?

there's a free tier — small wiki, sms-only, one connected ai. it's enough to see the shape. pricing →

15

what does pro give me?

larger vault, every connector, faster agent, per-user encryption at rest, and an export you actually own. monthly, cancel anytime. pricing →