install

Install Vertiso Memory for every tool.

Start with your tool's Marketplace when Vertiso Memory is published there. Until then, install the complete plugin from Vertiso's public GitHub repository when your tool supports it, or connect directly to our remote Model Context Protocol server with browser OAuth. Prefer the terminal, need a local bridge, or want to script your memory? The vmem CLI is further down.

1. Choose your AI tool

Each panel shows the best available route in order: a live Marketplace listing, the public Vertiso plugin repository, native HTTP MCP, then the local vmem bridge. Unpublished Marketplace links stay hidden until they work.

Vertiso Memory MCP URL 
https://memory.vertiso.ai/mcp

Pin your user card for the best experience.

Once Vertiso Memory has synthesized a knowledge card about you (automatic — happens after you've saved a few memories), pin vertiso-memory://knowledge/user/self in your client's resource picker. It's the single best piece of ambient context — your preferences, working style, recurring constraints — distilled and kept current automatically.

A page for each tool: Claude Claude Desktop Claude Code Cursor Codex ChatGPT VS Code JetBrains Raycast Antigravity

2.Your first conversation

Connecting only puts the vertiso-memory tools on the shelf — your assistant won't reach for them until it knows this is its memory. The standing instruction below works in every client, regardless of install path, and calls the hello tool at the start of each session. Full plugins also add checkpoint, handoff, handoff-resume, and wrap-up skills.

Standing instruction — paste into your client's custom-instructions field 
Vertiso Memory (vmem) is my persistent memory across every tool and session.

- Call the `hello` tool at the start of every session. Treat its primer
  as authoritative context, not a suggestion.
- `recall` before planning or answering anything that may depend on
  earlier context. Look it up instead of asking me to repeat it.
- `remember` durable decisions, preferences, constraints, and project
  state the moment they land. `update_memory` rather than writing a
  near-duplicate.
- Offer a handoff before the session ends, compacts, or moves to
  another tool.

Every client keeps this instruction somewhere different — the panel for your tool in step 1 says where, when it has a spot for one. Saved it (or on a client without one)? Start a new conversation and say Hello.

You'll know it worked when…

your assistant greets you by name and mentions what it remembers about you. On a brand-new account it simply confirms it registered — there's nothing to recall yet. From here, just talk to it: ask it to remember something, or recall what you told it earlier, and it reaches for Vertiso Memory on its own.

Want to shape what gets saved, organize your memories, or carry context across tools? Read the Configuring & using Vertiso Memory guide.

3.Power users: the vmem CLI

Optional for most tools, and required for the current JetBrains connection. The vmem CLI is a native binary (no Ruby/Node) for working memory from the terminal — scratch vmem recall, cron jobs, the live vmem watch view — and a stdio→HTTP MCP bridge for clients that can't reach a remote server yet (Antigravity 1), or when you'd rather not run the browser OAuth flow. It keeps your bearer token in ~/.vmem instead of in each app's config.

Install vmem 
curl -fsSL https://memory.vertiso.ai/install.sh | sh

Already installed the complete plugin for a client?

Do not also run vmem mcp install for that client unless you intend to replace its plugin-managed setup. The CLI installer can write another MCP server entry, skill copies, and a user-scoped SessionStart hook. For Claude, installing vmem and keeping it on PATH is enough for the plugin-bundled hook to run. Also running vmem mcp install claude-code can create duplicate setup.

Common commands

Run vmem with no arguments, or vmem <command> --help, for the full list.

Connect an MCP client

vmem mcp install <client> Wire Vertiso Memory into a client (cursor, vscode, claude-code, codex, antigravity, ...). No arg opens a picker.
vmem mcp list List the clients vmem can configure.
vmem mcp status Show which clients are already wired.
vmem mcp remove <client> Remove the Vertiso Memory entry from a client.
vmem mcp upgrade Migrate wired clients to the current bridge.
vmem mcp serve --client <slug> The stdio bridge a client launches; rarely run by hand.

Work with memories

vmem remember "<text>" Save a new memory.
vmem recall "<query>" Ask a question; get a synthesized answer.
vmem search "<query>" Ranked records, no synthesis.
vmem upload <path> Ingest a local file (PDF, image, ...) as a memory.
vmem ls List recent memories.
vmem show <id> Show one memory by id.
vmem rm <id> Delete a memory.

Sessions & setup

vmem init Save your API URL + token to ~/.vmem.
vmem whoami Verify the connection and your identity.
vmem hello Register the agent; print the session primer.
vmem handoff · handoff-resume Write or pick up a session handoff.
vmem import Scan known paths and import memory files.

API token (for the CLI and bearer-token fallbacks)

Connecting via OAuth above? You don't need this — you sign in during the connector handshake instead. The vmem bridge and the bearer-token fallback snippets authenticate with a token tied to your account; generate one and copy it to your password manager — it's shown once.

Log in to get your token →