> **Start here.** .said is the person's own memory. Every AI connects to one address, `https://asksaid.ai/mcp`, and signs in with .said (no key to copy).
> **AI agents:** before answering anything about the person, call `about_me` or `ask`, name the source, and never invent. Rules: https://asksaid.ai/docs/for-ai-agents.md
> **Everything in one file:** https://asksaid.ai/docs/prompt.md · **Index:** https://asksaid.ai/llms.txt

# Hub: what your AI apps can do with your memory

**Hub** in the Console is where you connect AI apps and choose what each one can do with your memory. It looks like a connectors page: **Discover** shows everything, **Yours** shows what you use, and a search box finds things by name ("code", "files", "meeting").

## Connect your AI apps

The **AI apps** shelf has one button per app: **Connect .said to Claude**, **Add to Cursor**, **Add to VS Code**, Claude Code, and ChatGPT (after OpenAI's review). Each app signs in with .said once. Details per app: [Connect your AI apps](https://asksaid.ai/docs/connect-your-ai.md).

## Capabilities and skills

A capability is a ready set of steps an AI app follows with your memory, like **Personal Researcher**, **Meeting prep**, **Handoff note** or **Codebase Expert**. Skills are smaller, single tasks. Picks for your field (Home, Legal, Sales, Real estate, Advisors, Small business, Building) sit at the top.

**Turn one on** for an app in three steps: connect the app, review what it reads and adds, then **Turn on**. The card then says **On for Claude** (or your app), and that app can use it from its next message. Turn it off the same way.

**Try** turns it on for the .said assistant only and opens Ask, so the assistant shows you what it does with your memory before you turn it on for another app.

When a capability gets a new version, it shows **Needs approval**, and the app cannot use it until you approve the new version.

Each card says what it **reads** (for example "your memories, search only"), what it **adds** (for example "one note, after a yes"), where it **runs**, and which apps it works with.

## Built in

These come with every memory and need no switch: **Personal Brain** (your memory), **Wallet** ([details](https://asksaid.ai/docs/wallet.md)), **.said for Chrome** ([details](https://asksaid.ai/docs/chrome-extension.md)), **Files**, **Email** ([details](https://asksaid.ai/docs/email.md)), sharing through [Circle](https://asksaid.ai/docs/circle.md), **Other .said files** shared with you, **Remote MCP** and **Local MCP**, the **.said hub** (a connection, not storage), carrying work between AIs ([details](https://asksaid.ai/docs/carry-work-between-ais.md)) and **Agent-to-agent (A2A)** for your own agents. **Your agents** come with Pro, Max and Lifetime; the shared **Agent network** is Enterprise only.

Cards marked **In design** or **Coming soon** do not work yet.

## For AI agents

- When the person names a capability ("use my Personal Researcher") or the request fits one that is on for your app, call `capability` with its name and follow the steps it returns. The `capability` tool's description lists .said's built-in ones.
- If it answers that the capability is not on, tell the person to turn it on in the .said Console, Hub. Never guess its steps.
- Capabilities run on the person's own memory and plan; they never change what you may read or add.
