MnemoverseDocs

Appearance

Memory for Coding Assistants ​

Give your AI coding assistant a memory that outlasts the session, the tool and the machine.

The Problem ​

You told Claude Code your deploy target is Railway. Tomorrow, in a new session, it suggests Heroku: the decision stayed in yesterday's conversation and never reached CLAUDE.md or Claude Code's auto memory. You explained the database schema to Cursor last week, and today Claude Code asks about it again.

The Solution ​

One command adds the memory tools. With the memory rules in its instructions, your coding assistant saves what you settle and reads it back in later sessions and in other tools.

Claude Code — add via CLI:

bash
claude mcp add mnemoverse -s user \
  -e MNEMOVERSE_API_KEY=mk_live_YOUR_KEY \
  -e MNEMOVERSE_API_URL=https://core.mnemoverse.com/api/v1 \
  -- npx -y @mnemoverse/mcp-memory-server@latest

On Windows (PowerShell), paste the same command as one line — PowerShell does not read the \ line continuations:

powershell
claude mcp add mnemoverse -s user -e MNEMOVERSE_API_KEY=mk_live_YOUR_KEY -e MNEMOVERSE_API_URL=https://core.mnemoverse.com/api/v1 -- npx -y @mnemoverse/mcp-memory-server@latest

For detailed setup per tool, see the integration guides:

Or give your agent this URL to self-configure: https://mnemoverse.com/docs/llms.txt

What Gets Remembered ​

With the memory rules in its instructions, your coding assistant remembers:

CategoryExample
Project setup"Uses PostgreSQL + Prisma, deployed on Railway"
Preferences"Always use pnpm", "Prefers Tailwind over CSS modules"
Decisions"Chose REST over GraphQL for caching simplicity"
Lessons"Never deploy on Fridays — learned the hard way"
People"Alice owns the design system, Bob reviews API changes"
Patterns"Exponential backoff fixed timeout issues in this service"

Cross-Tool Memory ​

Write a memory in Claude Code → read it in Cursor. Any key or sign-in from the same Mnemoverse account reaches the same memory.

Monday in Claude Code:
  You: "We always use feature branches, never push to main directly"
  Claude: [stores memory]

Wednesday in Cursor:
  You: "Push this change"
  Cursor: [reads memory] "Created feature branch. PR to main as usual."

Make your assistant use it ​

Connecting Mnemoverse isn't enough — the agent has to actually call the memory tools. The reliable way is a standing instruction in your project's agent config (CLAUDE.md, .github/copilot-instructions.md, AGENTS.md, .cursor/rules).

→ Make Your Agent Use Memory — the copy-paste instruction block, the right file per tool, and how to verify it's working.

Working in several repositories with one account? Give each repository its own domain, such as project:acme-api, in its own CLAUDE.md or AGENTS.md, so one project's facts stay out of another's sessions: Keep each project's memories in its own domain.

Team Memory ​

Each engineer keeps their own API key; the team shares knowledge through a room (Beta) — a membership-checked shared memory pool addressed with domain: "xroom:<room_id>" on the same write/read calls:

  • A new team member's assistant can read the conventions the room already holds
  • Decisions saved in one person's session can be read in everyone's

Personal preferences stay in your own memory (domain: "user:{name}"); shared project context goes to the room. Usage is billed to the authenticated caller, never the room owner — and don't share one API key across the team: keys identify a person and are revoked individually (see Security).

Common questions ​

Why does my AI coding assistant forget my project between sessions? ​

Each session starts with a fresh context window. What carries over is what the assistant loads at startup: instruction files such as CLAUDE.md, AGENTS.md or .cursor/rules, and in Claude Code its auto memory, which Anthropic's memory documentation describes as per repository and "machine-local" (read 2026-10-11). Mnemoverse adds a memory store that lives in your account and is reached over the Model Context Protocol (MCP), so project setup, decisions and preferences the assistant saves can be read in later sessions, in other tools and on other machines.

How do I give Claude Code or Cursor persistent memory? ​

Add the Mnemoverse MCP server: one command for Claude Code (claude mcp add mnemoverse -s user -e MNEMOVERSE_API_KEY=... -- npx -y @mnemoverse/mcp-memory-server@latest), or a small JSON config for Cursor / VS Code / Devin Desktop. The assistant then calls memory_read before answering and memory_write after learning something.

Can my coding assistant remember context across different tools? ​

Yes. One Mnemoverse account is one memory. Write a memory in Claude Code and read it in Cursor: every MCP-capable tool and any REST client connected to the same account reads the same memories. ChatGPT's MCP connector signs in through the browser (OAuth) instead of taking the key, and sees the same memory when you sign in with the Mnemoverse account the key belongs to.

How do I make my coding assistant actually use its memory? ​

Add a short instruction block to your project's CLAUDE.md (or .github/copilot-instructions.md): call memory_read before answering questions about preferences, past decisions, or setup, and memory_write after learning something worth keeping. This is the most reliable activation lever.

Is there a free way to add memory to my coding assistant? ​

Yes — Mnemoverse has a free tier with no credit card required. Get a key at console.mnemoverse.com and add the MCP server in about 30 seconds.

Get Started ​

  1. Get an API key (free, 30 seconds)
  2. Pick your tool: Claude | Editors | ChatGPT
  3. Make your agent actually use it — add the standing-instruction block
  4. Start coding: the assistant saves what you settle and reads it back in later sessions