Imagine taking your memory from ChatGPT to Claude to Gemini, and bringing that same context into Claude Code, Codex, Cursor, and Google Antigravity.
Living Brain is an interoperable, efficient memory layer for AI. Connect through MCP or build with the API. Keep the context. Choose the model. Carry the work forward.
🧠 Persistent memory · 🔌 MCP + API · ⚡ Up to 22.6× token savings on LongMemEval S*
Create your Living Brain → · Connect with MCP → · See reported results ↓
Living Brain · Performance · Developer guide · The ecosystem · Explore a scenario · Open source · Follow us
You have already explained the project. Made the decisions. Shared the context. That knowledge should keep working for you wherever you work next.
The Living Brain gives AI a persistent memory layer that can travel across models and workflows. It also powers ChatChat, our messaging-first platform for proactive AI agents.
ChatGPT · Claude · Gemini · Claude Code · Codex · Cursor · Google Antigravity
The idea is simple: choose the AI experience you want, and connect your Living Brain through the client's supported MCP or API integration. Your memory becomes a resource you can bring into the workflow.
| How you want to work | Where to start |
|---|---|
| Connect through MCP with sign-in | Follow the MCP guide for connection and authentication instructions. |
| Use MCP with API credentials | Follow the MCP guide for API-based access. |
| Build memory into your own application | Start with the Quickstart and TypeScript SDK. |
| Work from your terminal | Explore the Living Brain CLI. |
Connection options depend on the host application's support. See the MCP documentation for setup details.
A project. A research topic. Your team's knowledge. Your next big idea.
Create a Living Brain around the context you want to carry forward, then bring it into your AI workflows. Proprietary memory scoring is central to how Living Brain approaches useful, efficient memory.
| Capability | Why it matters |
|---|---|
| Interoperable memory | Bring persistent context into different AI tools through MCP and API integrations. |
| Proprietary memory scoring | A dedicated approach to scoring memory, built into the Living Brain memory layer. |
| Connected context | Connect people, information, events, and goals. |
| Continuity across work | Carry relevant knowledge beyond a single conversation. |
| Developer access | Work through MCP, the API, the TypeScript SDK, or the CLI. |
Long-running work needs memory that is useful and efficient. Living Brain's focus is to carry relevant context forward while reducing the token burden of long-context workflows.
| Team-reported result | Evaluation |
|---|---|
| Up to 22.6× token savings | Long-context evaluation using LongMemEval S. |
| #1 on LoCoMo | Ranking reported by the Living Brain team. |
| #1 on LongMemEval | Ranking reported by the Living Brain team. |
*These are team-reported results. Public result links, the ranking's comparison set, and the token-savings baseline and configuration are needed to independently assess them. Results can vary by model and workload.
Coming soon: BrainBench, our self-hosted benchmark suite. Put Living Brain to the test in your own environment.
Create a Brain. Connect your tools. Start building with memory.
| Your starting point | Official documentation |
|---|---|
| Create your first Living Brain | Quickstart → |
| Connect through MCP | MCP, sign-in, and API access → |
| Explore workspace search | Search Pages API reference → |
| Build with TypeScript | TypeScript SDK → |
| Work from the command line | CLI → |
Browse all Living Brain documentation →
| Role | Explore | |
|---|---|---|
| 🧠 Living Brain | The memory and context infrastructure powering ChatChat. | Explore Living Brain → |
| 💬 ChatChat | A messaging-first platform for creating and working with proactive AI agents, bringing conversations, knowledge, tools, and memory together. | Discover ChatChat → |
| 🛠️ ChatAndBuild | The team behind the Living Brain | Learn More → |
Our belief: anyone can chat and build. Our mission is to make advanced intelligence more accessible, practical, and empowering.
Explore the possibilities. These scenarios illustrate the workflows we are building towards; connection details depend on your chosen tools.
💬 Plan in ChatGPT. Pick up the idea in Claude. Explore it with Gemini.
Imagine connecting the same Living Brain as you move between AI tools. The project brief, decisions, and relevant knowledge become context you can bring to the next conversation.
Choose the model for the task. Keep your Brain with you.
⌨️ Bring project memory into Claude Code, Codex, Cursor, or Antigravity.
Your project has a history: requirements, architectural decisions, constraints, and lessons learned. Imagine bringing that context into your coding workflow through a connected Living Brain.
Give your next coding session a head start.
🛠️ Build your own application with a Living Brain.
Choose your model and build a workflow around persistent memory. Start with the Quickstart, explore the TypeScript SDK, and consult the Search Pages API reference for workspace search.
Your application. Your choice of model. A Brain to carry context forward.
Memory is one part of continuity. Identity and ownership are another.
Non-Fungible Agents began as a framework for bringing identity, ownership, memory, and continuity to autonomous AI agents. The project is open source and formalised through BAP-578, a standard for intelligent digital entities with autonomous capabilities and verifiable ownership.
- Explore the Non-Fungible Agents repository →
- Read the BAP-578 specification →
- Browse our public repositories →
For setup, licensing, contribution guidance, and security reporting, consult the relevant repository.
Explore the published stories about ChatAndBuild:
We welcome builders, researchers, founders, and partners exploring what persistent intelligence can make possible.
Choose your AI. Bring your Brain. Keep building.