> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fwdslash.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Create your first agent

> Set up your first AI agent by defining its role, behavior, and knowledge, then train it to start responding.

## Sign up to FwdSlash

Create your FwdSlash account, either by providing an email address and password or with Google.

​On signing up, you will be guided through agent creation.

<Note>
  You'll need an active FwdSlash account to follow this guide. [Sign up](https://app.fwdslash.ai/signup) here if you haven't already.
</Note>

## Select your role

<Frame>
  <img src="https://mintcdn.com/techopsfwdslashai/wZiC_f8V3rmA3GFk/images/Select-your-role.png?fit=max&auto=format&n=wZiC_f8V3rmA3GFk&q=85&s=901f951640924c61f57cdda41cc72f9f" alt="Select Your Role" width="2880" height="1620" data-path="images/Select-your-role.png" />
</Frame>

You will see:

> "Which role best describes what you do?"

Options include:

* Marketing
* Product Management
* Research
* Growth
* Customer Support
* Other

This personalizes your setup. It does not affect your agent’s behavior directly.

## Configure your agent

<Frame>
  <img src="https://mintcdn.com/techopsfwdslashai/wZiC_f8V3rmA3GFk/images/Configure-your-agent.png?fit=max&auto=format&n=wZiC_f8V3rmA3GFk&q=85&s=389aef2c109ef05f5f048d1b77562917" alt="Configure Your Agent" width="2880" height="1620" data-path="images/Configure-your-agent.png" />
</Frame>

​Here, you set up your agent’s role, name, and behavior to define how it thinks and responds from the start.

You’ll define:

* Role → what the agent does
* Name → how it’s identified
* System Prompt → instructions that control behavior

Some fields are pre-filled based on your earlier selections to help you get started quickly, but you can:

* Edit everything freely
* Switch to a custom role at any time
* For custom roles, click Generate Prompt to automatically generate based on the name given

## Add knowledge (Optional)

<Frame>
  <img src="https://mintcdn.com/techopsfwdslashai/wZiC_f8V3rmA3GFk/images/Add-knowledge.png?fit=max&auto=format&n=wZiC_f8V3rmA3GFk&q=85&s=aa4d3803b42c8958121a5cb809b5a339" alt="Add Knowledge" width="2880" height="1620" data-path="images/Add-knowledge.png" />
</Frame>

You can provide:

* Website content
* Files
* Text blocks
* Q\&A pairs

This step improves response quality but is not required to finish creating your first agent.

<AccordionGroup>
  <Accordion title="Website">
    Add a website by entering a URL and choosing how to ingest it—crawl links, use a sitemap, or a single page. You can optionally refine this with include or exclude paths, then click Add.
  </Accordion>

  <Accordion title="Files">
    Upload files by dragging and dropping them into the upload area. Supported formats include PDF, DOC/DOCX, and TXT. PDFs must contain selectable text to be processed.
  </Accordion>

  <Accordion title="Text">
    Add a text block by providing a title and the corresponding body text.
  </Accordion>

  <Accordion title="Q&A">
    Add a Q\&A by entering a question and its answer, then save it.
  </Accordion>
</AccordionGroup>

Once you're done, hit Train Agent for the knowledge to take effect.

## Complete setup

You’ll see a View your agent in action button, which opens your agent in Chatground where you can interact with it and continue refining its behavior.