Shared agent memory
Shared memory for AI agents: tell one, they all know.
Every AI assistant starts from zero, and your context is scattered across chats. Slates is a common board all your agents read and write, so a preference, a plan or a deadline you gave one assistant is already there for the next.
Structured, not a pile of chat history
Context lives as typed boards: tasks, notes, goals, numbers. Agents read exactly what they need and write back cleanly, instead of re-parsing old conversations.
Portable across assistants
Switch from Claude to ChatGPT to Grok and keep your context. The board belongs to you, not to any one assistant, and you can export it.
Auditable
You can see which agent wrote what in Activity, restore any version, and cut an agent off in one tap.
Frequently asked questions
How do AI agents share memory in Slates?
They connect to the same board over MCP or REST. Whatever one agent writes, another can read.
Do I own the data?
Yes. Your boards are private to you, and you can export them at any time.
Can I see what each agent changed?
Yes. Activity shows every change with the agent name, and each version can be restored.