# Multitude > The shared memory layer for AI agents. One memory for Claude, Claude Code, Cursor, Codex and the agents your team runs, kept as plain Markdown your team owns. Every AI, briefed. Multitude is for founders running several AIs, and small teams of up to about 15, where Claude, Claude Code, Cursor and Codex need the same approved decisions. It keeps memory as Markdown on your own computers with end-to-end encrypted sync, separates suggestions from decisions, records who authorised each fact, flags contradictions and keeps replaced facts in history. Status: preview; the beta opens on 1 December 2026. Pricing: Free on one computer; Pro $19 a month (sync, decision history, approved memory, auto-recall); Team $99 a month billed yearly for the first 5 people, then $16 a seat ($119 plus $20 a seat monthly); founding teams $9 a seat for life (first 50 teams); Enterprise through a design partner program, priced by conversation. People, and the AIs they connect themselves, are never metered. Multitude never trains on customer data and is working towards SOC 2 Type I (not certified). ## Product - [Home](https://trymultitude.com/): Every AI, briefed. Say it once and every AI remembers. Claude, Claude Code, Cursor and Codex work from the same approved decisions, for founders and small teams. - [Teams](https://trymultitude.com/teams.html): Start on your own. When a contractor or teammate joins, share one project with them, each with its own encryption key. - [Pricing](https://trymultitude.com/pricing.html): Free on one computer. Pro $19 a month with sync, decision history and auto-recall. Team $99 a month for 5 people then $16 a seat, billed yearly. Enterprise through our design partner program. People are never metered. - [Security](https://trymultitude.com/security.html): Plain Markdown you own, permissions that follow the source, an audit log of every AI read, and no training on your data. - [Integrations](https://trymultitude.com/integrations.html): Claude Code, Codex, Cursor, Windsurf, VS Code and Gemini CLI in one command. Granola, Slack, Notion, Drive, GitHub, Linear, Zoom and more as notes. - [Docs](https://trymultitude.com/docs.html): Multitude docs for the shared memory layer for AI agents: get started, set up a team, connect Claude, Claude Code, Cursor and Codex, connect work tools, privacy and sharing. - [FAQ](https://trymultitude.com/faq.html): Answers about Multitude: what it is, which AIs and tools it works with, where data lives, pricing, privacy and when the beta opens. ## Guides - [What is shared AI memory?](https://trymultitude.com/shared-ai-memory.html): What shared AI memory is, why ChatGPT and Claude memory fall short for teams, and the five things to look for: current facts, authority, permissions, portability and every AI. - [Glossary](https://trymultitude.com/glossary.html): Plain definitions of AI memory terms: shared AI memory, MCP, context engineering, supersession, provenance, AGENTS.md and more. - [Hand work from Codex to Claude Code](https://trymultitude.com/playbook-codex-to-claude-code.html): A step-by-step playbook for passing context between OpenAI Codex and Claude Code with one shared Markdown memory and session hooks. - [Turn meetings into team memory](https://trymultitude.com/playbook-meeting-to-memory.html): Granola, Zoom, Fireflies or Fathom notes become Markdown your Claude, ChatGPT and Gemini can search, with decisions confirmed by a person. - [Brief every AI on a new client](https://trymultitude.com/playbook-client-onboarding.html): An agency playbook: one shared client folder with brand, tone, contacts and decisions that every teammate's AI reads. - [Switch between ChatGPT and Claude](https://trymultitude.com/playbook-switch-ai.html): Export your ChatGPT history, import it as Markdown and keep one memory that works in ChatGPT, Claude and Gemini. - [Brief a new hire's AI on day one](https://trymultitude.com/playbook-new-hire.html): Share your team folders with a new teammate and their Claude, Claude Code or Cursor knows your clients, decisions and conventions from the first prompt. - [Keep decisions current across every AI](https://trymultitude.com/playbook-decision-log.html): Stop AIs quoting outdated prices and abandoned plans. How Multitude separates suggestions from decisions and keeps the history. ## Use cases - [Agencies](https://trymultitude.com/for-agencies.html): Give every person and every AI at your agency the same client context: brand rules, decisions, contacts and history, in Claude, Claude Code, Cursor and Codex. - [Startups](https://trymultitude.com/for-startups.html): Startups use Claude, Claude Code, Cursor and Codex side by side. Multitude gives them all the same product decisions, pricing and customer context. - [Design studios](https://trymultitude.com/for-studios.html): Designers in Claude, engineers in Cursor, account leads in ChatGPT: one team memory for every project. How studios use Multitude. - [Consultancies](https://trymultitude.com/for-consultancies.html): Keep engagement context, interview notes and recommendations in one permissioned memory every consultant's AI can use. Built for small consultancies. - [Developers](https://trymultitude.com/for-developers.html): The shared memory layer for AI agents. Hand work between Claude Code, Codex and Cursor without re-explaining. Local MCP memory with session hooks, plain Markdown and no API keys. ## Comparisons - [Multitude vs Basic Memory](https://trymultitude.com/vs-basic-memory.html): Basic Memory vs Multitude: both keep AI memory as Markdown. How they differ on team pricing, decisions and contradictions, and which AI apps they reach. Checked 21 November 2026. - [Multitude vs Mem0](https://trymultitude.com/vs-mem0.html): Mem0 is a memory API for developers building AI apps. Multitude gives your team's Claude, Claude Code and Cursor one shared memory. Plans, limits and when to choose each. - [Multitude vs Supermemory](https://trymultitude.com/vs-supermemory.html): Supermemory is a memory and retrieval API priced by usage. Multitude is a shared Markdown memory for teams at a flat price. Honest comparison, checked 21 November 2026. - [Multitude vs Letta](https://trymultitude.com/vs-letta.html): Letta (formerly MemGPT) is a platform for building stateful agents. Multitude gives the AI tools your team already uses one shared memory. Plans and differences, checked 21 November 2026. - [Multitude vs Pieces](https://trymultitude.com/vs-pieces.html): Pieces captures your own work into on-device long-term memory. Multitude gives a whole team one shared Markdown memory across every AI. How they differ, checked 21 November 2026. - [Multitude vs Obsidian](https://trymultitude.com/vs-obsidian.html): Obsidian is a free, local Markdown notes app. Multitude makes an Obsidian vault readable by Claude, Claude Code, Cursor and Codex, and shareable with a team. Use them together. - [Multitude vs ChatGPT Memory](https://trymultitude.com/vs-chatgpt-memory.html): ChatGPT Memory remembers you inside ChatGPT. Multitude is one memory every AI shares, kept as notes your team owns. What each does, checked against OpenAI's help centre on 21 November 2026. - [Multitude vs Claude Memory](https://trymultitude.com/vs-claude-memory.html): Claude Memory remembers you inside Claude, per project. Multitude gives Claude, Claude Code, Cursor and Codex one shared memory a team can own. Checked against Anthropic's help centre, 21 November 2026. - [Multitude vs Notion AI](https://trymultitude.com/vs-notion-ai.html): Notion AI works inside Notion on Business and Enterprise plans. Multitude brings your team's memory into Claude, Claude Code, Cursor and Codex. Differences and how to use both. - [Multitude vs CLAUDE.md and AGENTS.md](https://trymultitude.com/vs-claude-md-agents-md.html): CLAUDE.md and AGENTS.md brief coding agents from inside a repository. When a file is the right tool, where it stops, and how Multitude fits alongside it. ## Integrations - [Claude](https://trymultitude.com/integrations-claude.html): Connect Claude to Multitude: give Claude on the web, desktop and mobile your team's shared memory. Shared Markdown memory that Claude, Claude Code, Cursor and Codex all read. - [ChatGPT](https://trymultitude.com/integrations-chatgpt.html): Connect ChatGPT to Multitude: let ChatGPT read the same memory as Claude, Gemini and your coding agents. Shared Markdown memory that Claude, Claude Code, Cursor and Codex all read. - [Gemini](https://trymultitude.com/integrations-gemini.html): Connect Gemini to Multitude: bring your shared memory into Gemini. Shared Markdown memory that Claude, Claude Code, Cursor and Codex all read. - [Claude Code](https://trymultitude.com/integrations-claude-code.html): Connect Claude Code to Multitude, the shared memory layer for AI agents. Brief Claude Code at the start of every session, and save what it learned at the end. Plain Markdown, local MCP, no API keys. - [Codex](https://trymultitude.com/integrations-codex.html): Connect Codex to Multitude, the shared memory layer for AI agents. Give OpenAI Codex the same memory as your other agents. Plain Markdown, local MCP, no API keys. - [Cursor](https://trymultitude.com/integrations-cursor.html): Connect Cursor to Multitude, the shared memory layer for AI agents. Cursor, briefed on your product, clients and decisions. Plain Markdown, local MCP, no API keys. - [Windsurf](https://trymultitude.com/integrations-windsurf.html): Connect Windsurf to Multitude, the shared memory layer for AI agents. Connect Windsurf to your shared memory over MCP. Plain Markdown, local MCP, no API keys. - [VS Code](https://trymultitude.com/integrations-vscode.html): Connect VS Code to Multitude, the shared memory layer for AI agents. Use your shared memory from agents in VS Code. Plain Markdown, local MCP, no API keys. - [Gemini CLI](https://trymultitude.com/integrations-gemini-cli.html): Connect Gemini CLI to Multitude, the shared memory layer for AI agents. Give Gemini CLI the same memory as Claude Code and Codex. Plain Markdown, local MCP, no API keys. - [Granola](https://trymultitude.com/integrations-granola.html): Sync Granola into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Wispr Flow](https://trymultitude.com/integrations-wispr-flow.html): Sync Wispr Flow into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Notion](https://trymultitude.com/integrations-notion.html): Sync Notion into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Slack](https://trymultitude.com/integrations-slack.html): Sync Slack into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Linear](https://trymultitude.com/integrations-linear.html): Sync Linear into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [GitHub](https://trymultitude.com/integrations-github.html): Sync GitHub into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Google Drive](https://trymultitude.com/integrations-google-drive.html): Sync Google Drive into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Inbox (email and webhooks)](https://trymultitude.com/integrations-inbox.html): Sync Inbox (email and webhooks) into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Gmail and Outlook](https://trymultitude.com/integrations-gmail.html): Sync Gmail and Outlook into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Zapier, Make and n8n](https://trymultitude.com/integrations-zapier.html): Sync Zapier, Make and n8n into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Zoom](https://trymultitude.com/integrations-zoom.html): Sync Zoom into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Fireflies](https://trymultitude.com/integrations-fireflies.html): Sync Fireflies into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Fathom](https://trymultitude.com/integrations-fathom.html): Sync Fathom into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Google Calendar](https://trymultitude.com/integrations-google-calendar.html): Sync Google Calendar into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Jira](https://trymultitude.com/integrations-jira.html): Sync Jira into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Confluence](https://trymultitude.com/integrations-confluence.html): Sync Confluence into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Asana](https://trymultitude.com/integrations-asana.html): Sync Asana into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. - [Readwise](https://trymultitude.com/integrations-readwise.html): Sync Readwise into plain Markdown notes that Claude, Claude Code, Cursor and Codex can read. Read-only, permission-aware, in the Multitude beta. ## Optional - [Full facts in one file](https://trymultitude.com/llms-full.txt) - [Subprocessors](https://trymultitude.com/subprocessors.html) - [Privacy](https://trymultitude.com/privacy.html) - [Terms](https://trymultitude.com/terms.html)