> ## 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.

# How Superscale works

> How research, creation, and media buying hand work off to each other.

Superscale is organized around a research -> creation -> media-buying loop. Research gives the agent market and account signal. Creation turns that signal into scripts, statics, videos, and variations. Media buying and reporting show what happens after launch, so useful findings can be saved back into context.

## The three stages

<Frame>
  <img src="https://mintcdn.com/superscale/Pw84ChmqHn6vjQyU/images/Vision.png?fit=max&auto=format&n=Pw84ChmqHn6vjQyU&q=85&s=cad35c61a8f04cf717a7aa31bc9ae445" alt="Vision" width="1440" height="457" data-path="images/Vision.png" />
</Frame>

## What each stage does

Read the loop as a hand-off. The stages do not all run on their own; each stage takes inputs, produces a working artifact, and may create a learning worth saving as context.

| Stage                                                | Uses                                                                                            | Produces                                                                          | Cost                                                                          | What carries forward                                                              |
| ---------------------------------------------------- | ----------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------- | ----------------------------------------------------------------------------- | --------------------------------------------------------------------------------- |
| [Research & analysis](/features/competitor-research) | Brand and product context, competitor ads, connected account data, reviews, and public research | Reports, scored ads, competitor lists, creative briefs, and hypotheses            | Usually lower than production; reports and messages use credits               | Findings become inputs for briefs, references, or context files                   |
| [Creation](/features/static-ads)                     | A brief, brand and product context, exact assets, and visual references                         | Static ads, video ads, scripts, resized exports, and iterations                   | Most credit use lives here; static and video generation have different costs  | Strong outputs, corrections, and winners can become references or context         |
| [Media buying](/how-it-works/media-buying)           | Approved creative, connected ad accounts, platform permissions, and campaign prerequisites      | Manual exports, supported platform writes, performance reads, and recommendations | Ad spend is external to credits and stays under your platform account control | Performance results and recommendations guide the next research or creative brief |

<Note>
  Autonomous does not mean unattended platform action. Ad-platform writes that can affect delivery, spend, budget, or live creative wait for approval before the gated action is sent.
</Note>

The loop improves when you deliberately keep the useful parts: save durable findings as context, keep strong ads as references, and turn repeated corrections into rules.