> ## Documentation Index
> Fetch the complete documentation index at: https://docs.superscale.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Context, explained
> What context is, why it decides the quality of your ads, and what to put in it.
Context is the durable source material Superscale can reuse: your brand, products, customers, claims, assets, and market learnings. It is the foundation for accurate, on-brand creative.
Here's the core idea: **context reduces guessing.** When the agent writes a hook or picks an angle, it can use stored facts, customer language, reference assets, competitor findings, and your own rules instead of relying only on the current prompt.
## Why context-first
Prompt-only workflows make each new ad depend on the current brief. Marketing usually needs repeated output across the same brand, products, audiences, and rules, so the same facts should not have to be rewritten for each task.
Superscale moves that effort upstream. Set up context once, keep it current, and the agent can reuse it whenever you create for that workspace, brand, or product.
## It lives at three levels
Context is nested. Anything you set up can apply to your whole **workspace**, a single **brand**, or one **product**, and it cascades down.
What the agent uses now, what stays temporary, and what becomes durable context.
Where context lives, and where each piece comes from.
## The two kinds of context
Your context comes together from two directions.
### What Superscale gathers for you
Start from a supported source and Superscale can create the first version of your brand and product context. What comes in depends on the source and what can be read from it:
Analyzed to build brand identity and product or offer details from the pages Superscale can read.
Product details and imagery imported from supported Shopify or Amazon sources, or added manually.
App Store and Google Play listings, screenshots, metadata, and reviews when those sources are provided.
Competitors and advertiser pages found through research flows, then reviewed before they become useful references.
Connected ad accounts and advertiser pages that can add campaign and performance signal once linked.
Public web and conversation research that can be used when you ask the agent to investigate a market, audience, or angle.
### What you add: context files
The rest is the knowledge only *you* have: the things that are not on your website and cannot be reliably inferred. You add these as **context files**: scoped documents, notes, and data files the agent can read.
Your website is your pitch. Context files are where you put internal knowledge: what customers really say, what legal will reject, what the brand will never do, and which lessons should survive beyond one chat.
Things worth putting in a context file:
* Your ICPs: who you're really selling to, their pains, and what makes them buy.
* Customer testimonials, the quotes that capture why people love you.
* Standout reviews, in the exact words your customers use.
* Internal insight: positioning, common objections, and lessons that never made it onto a page.
When the agent writes ad copy, it reads this to understand how your customers think and speak. The more of your real, non-public knowledge you give it, the less it falls back on generic phrasing.
## Set up your context