guide

Configuring & using Vertiso Memory.

You've connected Vertiso Memory and said your first Hello. Here's how to actually work with it — save what matters, get it back, keep it tidy, and carry it across every tool you use. Haven't connected yet? Start with the install guide.

1.How your memory works

Vertiso Memory is a store your AI assistant reads from and writes to as you work. You don't manage it like a filing cabinet — you just talk. When something is worth keeping, your assistant saves it; when something you've told it before is relevant, it pulls it back. The result is continuity: your tools stop forgetting.

At the start of a session, the hello step loads a short primer — who you are, your preferences, what you're working on, and any open handoffs — so the assistant is oriented before you type. That's why the standing instruction from the install guide matters: it tells your assistant to register on startup and treat Vertiso Memory as its memory.

Saving happens two ways, and both are normal: your assistant remembers durable facts on its own as you work, and you can always tell it to remember something explicitly. Ephemeral back-and-forth stays out unless you ask for it.

2.Saving what matters

To save something on purpose, just ask in plain language: "Remember that I prefer TypeScript over JavaScript," or "Save this decision: we're launching on the Connectors Directory." Good memories are durable and reusable — your preferences, decisions, project state, constraints, and goals. Skip the throwaway details; Vertiso Memory is for what you'll want again next week, not the current sentence.

Behind the scenes each memory has a type so it can be organized and recalled well. You rarely set these by hand, but it helps to know the shape of them:

decision A choice you've made, with the reasoning behind it.
project A named body of work you own: the container your actions roll up into, serving the goals it delivers on.
constraint Rules and guidance, graded by importance — from how you like to work to hard limits that must hold.
action A single task to get done — the to-do itself.
intent A goal with a measurable outcome — the "why" your projects and actions serve.

…plus observations, events, and lessons. Your assistant picks the right type for you.

You can also feed it files. Ask your assistant (or use vmem upload) to bring in a PDF, image, or note, and Vertiso Memory extracts it into memory you can recall later.

3.Getting it back

There are two ways to pull memory back, and they answer different questions:

recall Ask a question, get a synthesized answer with its sources and a confidence level. "What did we decide about pricing?"
search Get the matching records themselves, ranked, without synthesis — for when you want to see the raw memories.

Most of the time you won't invoke either by name — you'll just ask your assistant a question, and it recalls what it needs. Reach for an explicit "search my memories for…" when you want to browse rather than get an answer.

4.Keeping it tidy

Memory is easiest to trust when it's current. A few habits keep it that way:

  • Correct, don't duplicate. When something changes, tell your assistant to update the existing memory rather than adding a second, conflicting one.
  • Archive vs. delete. Archiving keeps a memory but marks it done or inactive, so it stays recallable without cluttering the present. Deleting removes it for good.
  • Tag what you'll group. Tags make related memories easy to pull together later.

Prefer to browse and edit directly? Signed in on the web, your Memories, Knowledge, and Intents views let you read, edit, archive, and delete everything by hand.

5.Your user card

Once you've saved a handful of memories, Vertiso Memory synthesizes a user card about you — your preferences, working style, and recurring constraints, distilled into one compact profile and kept current automatically. It's the single most useful piece of context you can hand an assistant.

Pin it for the best experience.

In your client's resource picker, pin vertiso-memory://knowledge/user/self so your card rides along as ambient context in every conversation — no re-explaining who you are.

6.Across tools & sessions

This is the payoff: your memory isn't locked to one app. Save a preference in Claude, and it's there when you open Cursor, ChatGPT, or the terminal. Connect Vertiso Memory in each tool once (see the install guide) and they all draw from the same memory.

For longer work that spans sessions, use a handoff. Ask your assistant to write one when you're wrapping up, and it records where things stand; next time, a resume picks it up so you continue exactly where you left off instead of re-briefing from scratch.

7.You're in control

It's your memory. Everything Vertiso Memory holds about you is visible and editable from the web app — review any memory, correct it, archive what's gone stale, or delete it outright. Nothing is locked in, and you decide what's worth keeping.

8.Good habits & going deeper

A few things that make Vertiso Memory pay off faster:

  • Say Hello to start a session — it loads your primer and sets the assistant up right.
  • Be explicit when it matters: "remember this," "update that," "what did we decide about…?"
  • Pin your user card everywhere you work.

Want to script your memory, wire it into automations, or build on top? The vmem CLI and the full API reference cover the developer surface — the CLI is in the Power-users section of the install guide.