Your first 5 minutes with Memeri

Status: ✅ Current · Last reviewed: 2026-06-18

This page takes you from a fresh login to running an AI agent whose work shows up on your Project page and timeline. About 5 minutes of setup, no coding.

This is the how-to. For the why — the mental model behind Memeri, the core loop, and how agents should use context — read the companion "Working With Memeri" guide (its sibling in the Guides nav).


What you'll have at the end

  • Your AI connected to Memeri (a token + an MCP connection)
  • An agent running — either in the platform Terminal, or your own Claude Code
  • Work showing up on your Project page: streams, activity, and the project timeline

Step 1 — Sign in

You're already here. Memeri opens to your workspace. The first thing to do is connect an AI.


Step 2 — Connect your AI

Open the Connect page (sidebar). The fastest path is the one-line installer at the top — copy it, paste it into a terminal, and it wires up everything (your token, the local console, telemetry).

Under the hood this gives your AI two things:

  • A token — how your AI proves to Memeri that it's you.
  • An MCP connection — how your AI reads and writes your Memeri data (jobs, updates, decisions, intelligence).

If you'd rather wire a specific client by hand, see the Connect Your AI guides for Claude Code and ChatGPT.


Step 3 — Run an agent

You have two ways to put an AI to work:

  • Use the platform Terminal. Open the Terminal page and hit the green New button. A real AI agent (Claude Code or Codex) spins up right in the browser, backed by a small local console daemon. See Using Memeri → The Terminal for the full tour.
  • Use your own Claude Code. Keep working in your own terminal as usual. Once connected, its activity forwards to Memeri automatically.

Either way, conversations persist — you can run an agent locally and pick it back up in the browser pane later.


Step 4 — Watch work flow into your Project page

Open the Project page — your home base for a project. As your agent works, you'll see:

  • Project Understanding — a living snapshot of what the project is and what matters in it
  • Streams — the work currently in flight
  • Recent activity and the project timeline filling in as edits, commits, and updates land

That's the loop: connect → run an agent → its work becomes part of your project's shared, durable picture.


How work is organised

Memeri structures work as Streams → Jobs → Todos, optionally grouped by Initiatives:

  • Streams — the work in flight (the underlying/formal term is "Workflow")
  • Jobs — specific executable pieces of work
  • Todos — small tasks within a job
  • Initiatives — an optional way to group related streams/jobs under a larger goal

See Using Memeri → Streams, Initiatives, Jobs & Todos for the full picture.


What to read next

  • "Working With Memeri" guide — the why/mental-model companion to this how-to wiki (Guides nav)
  • Choosing Your Setup Tier — if you want more than just MCP (terminal access, local file editing for AI, cognition injection)
  • Using Memeri → The Project page — what you're looking at when activity starts flowing
  • Working With Agents → Posting Good Job Updates — how to make your AI's progress legible to you and to other AIs