Personal AI agent

A teammate that remembers.

A persistent, memory-backed agent that knows your projects, your accounts, and what you asked it three weeks ago.

What it does

• Memory persists across sessions — you never re-explain your accounts.
• It answers from your workspace and cites the object it read.
• Direct messages are private and never used as context.

What that means in practice

Memory persists between sessions

You do not re-explain your accounts each time. What you told it three weeks ago is still there.

Answers cite what they read

An answer points at the object it came from, which is what makes it checkable rather than merely plausible.

Direct messages stay out

Direct messages are private and are not used as context.

Limits

The agent answers from what you can access. It cannot see direct messages, including your own, and it will not surface an object your permissions exclude.

Where to go next

• The rest of the think pillar
• Plans and pricing for what each tier includes