# Memory

> Upload brand guides and checklists once. Memory learns each client from every review, feeds the QA agents, and powers Ask AI. Your team decides.

This is the Markdown copy of https://usesuperflow.ai/memory, published for AI agents and scripts. The HTML page is canonical.

## Key facts

| Field | Value |
| --- | --- |
| Page type | Feature page |
| URL | https://usesuperflow.ai/memory |
| Feature | Memory |
| Promise | Superflow remembers every client |
| Capabilities documented | 4 |

## What this does

Upload brand guides and checklists once. Memory learns the rest from every review: rules, taste, past decisions. Agents check against it. Ask AI answers from it.

Every client, remembered

Memory is Superflow's per-client brain: it stores what you upload and learns what every review teaches.

## Capabilities

Superflow gets you from First Upload to Sharper Reviews.

| Capability | What it does |
| --- | --- |
| What goes in | Brand guides, guidelines, and checklists go in once, and every review teaches Memory more about each client. |
| What it holds | Everything Memory keeps is scoped to one client - rules, taste, and decisions, held apart from every other account. |
| What it feeds | Memory grounds the agents, powers Ask AI, and surfaces suggestions before mistakes ship. |
| Why it compounds | The more a client reviews, the more their next review starts already knowing. |

## Detail

- **What goes in: One-time uploads.** Brand guides, guidelines, and review checklists go in at setup, per client.
- **What goes in: Learning from reviews.** Every accept or reject teaches Memory more about that client.
- **What it holds: Per-client memory.** Rules, taste, and decisions stay scoped to one client, never bleeding into another.
- **What it holds: Scoped three ways.** Agency-wide standards, per-client rules, and any knowledge base you upload.
- **What it feeds: Agent grounding.** Agents check every new asset against that client's memory, not just generic rules.
- **What it feeds: Proactive suggestions.** Memory flags a likely miss before review starts, from what this client rejected before.
- **What it feeds: Ask AI's source.** Every answer Ask AI gives is grounded in what Memory holds.
- **Why it compounds: Sharper every project.** Project ten starts already knowing what project one had to learn.

## How to start

Four steps, no engineer required.

1. **Add the snippet in 30 seconds.** Or upload a file. One click for WordPress, Webflow, Framer, Shopify.
2. **Teach it once per client.** Upload brand guides, guidelines, and checklists. Memory loads them for every review.
3. **Review against that client's memory.** Agents check every asset against it, and every human comment teaches it more.
4. **Sign off, and it remembers.** What your client approves becomes the rule next time, so your team doesn't have to.

## Questions and answers

### What does Memory actually store?

Two things: what you upload - brand guides, guidelines, and checklists - and what reviews teach it: each client's rules, taste, and decisions, scoped per client.

### Is my data used to train models for other customers?

No. One client's memory never informs another client, and one customer's data never trains another's.

### Can I see and edit what it remembers?

Yes. Every entry is inspectable, editable, and deletable, and cites the upload or review that taught it.

### Is this the same brand guide I upload when building an agent?

Yes - one upload, one record. Teach it in Memory settings or during agent creation; both feed the same client memory.

### Does Memory work from day one, or does it need history?

Day one: uploads apply immediately. The learning half compounds from your first review onward.

### What happens when a client rebrands?

Replace the upload. New decisions supersede old entries, and agents check against the current record.

### How is this different from a shared drive of brand guides?

A drive stores documents. Memory applies them: agents check against it, Ask AI answers from it, and every review adds to it.

### What does Memory cost?

It's included on every paid plan - it powers the agents and Ask AI rather than billing separately.

## Related pages

Each of these also publishes a Markdown copy at its own path plus `.md`.

- [AI review agents](https://usesuperflow.ai/ai-review-agents) - The checks Memory makes client-specific.
- [Ask AI](https://usesuperflow.ai/ask-ai) - The questions Memory makes answerable.
- [Client review](https://usesuperflow.ai/client-review) - The approvals that teach Memory what each client accepts.

## About Superflow

Superflow is a website and creative-asset review tool. Its agents watch every page of a site, report what changed, and route findings to the people who approve them. Teams leave contextual feedback on live sites, staging environments, PDFs, images, videos, and Lottie animations, and comments sync two-way with Asana, ClickUp, Monday, Slack, Webflow, and Google Tag Manager.

Machine-readable entry points:

- Site index: https://usesuperflow.ai/llms.txt
- Full site content in one fetch: https://usesuperflow.ai/llms-full.txt
- Free tools, with a Markdown copy of each: https://usesuperflow.ai/tools.md
- MCP server (Streamable HTTP, no account and no API key): https://usesuperflow.ai/api/mcp
- Agent card: https://usesuperflow.ai/.well-known/agent-card.json