What it can do
Where the folder actually helps.
Query me does one thing: it keeps one durable, editable set of notes and hands the right one to whichever assistant is in front of you. What that is worth to you depends on the shape of your work. Below are the situations it tends to actually change, drawn from the mechanics on the how it works page, not from feature promises that go beyond them.
Where it applies
Four kinds of work it tends to change.
Freelance and client work
Each client gets one note: how they like to be written to, what was agreed on the last call, what is still open. Draft an email in ChatGPT on Monday and follow up from Claude on Thursday, and the tone, the history and the open items are the same both times, because both are reading the same file instead of two separate chat histories.
Research and long-running projects
A project that runs for months accumulates decisions, dead ends, and reasons things were ruled out. Query me keeps one current version of that record instead of it living across dozens of old conversations you would otherwise have to search or re-explain, and an archive/ note for what used to be true, kept only as long as it still explains a decision.
Small teams and shared work
The QM Team Pilot extends the same idea to more than one person: a shared note that has exactly one owner, so two people's assistants cannot quietly disagree about what was decided, and a handoff format so one person's assistant can leave a note for a colleague's. Scoped separately from the self-serve editions; see pricing.
Ordinary personal use
Standing preferences, recurring dates, people you correspond with, the small facts you are tired of repeating. This is the least glamorous case and the one nearly everyone who tries it actually uses first: an assistant that already knows how you like an email written, without a settings screen to maintain per app.
What it is not
Where it does not help.
It is a memory, not a workflow platform.
Query me stores and routes text; it does not connect to your calendar, inbox, CRM, or other apps, and it does not automate tasks on its own. It works only inside the one folder you set up for it, and only with an AI surface that has permission to reach that folder, so it will not help with a tool that cannot connect to Google Drive.
It also will not make an assistant more capable at reasoning, writing, or coding. It changes what an assistant already knows about your situation going in, not how well it performs once it knows it.
See what has actually been tested.
The scenarios above describe what the mechanism is designed to do. The evidence page separates that from what has been measured so far.