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

# Playwright

> Let your UnleashX agent navigate pages, click and fill forms, screenshot, and extract content from any website.

# Playwright

> **In one line:** Add Playwright so your UnleashX agent can open web pages, click and fill forms, take screenshots, save pages as PDF, and pull text and content from the live web — all in plain language.

|                    |                                                                                                                                        |
| ------------------ | -------------------------------------------------------------------------------------------------------------------------------------- |
| **Category**       | Dev Tools & Automation                                                                                                                 |
| **Authentication** | None (built-in hosted browser)                                                                                                         |
| **Setup time**     | \~1 minute                                                                                                                             |
| **Difficulty**     | Easy                                                                                                                                   |
| **Best for**       | Teams that need an AI agent to read live web pages, interact with sites, and capture content without setting up any browser or account |

## 1. Overview

Playwright is a browser automation engine that drives a real, headless web browser — navigating pages, clicking elements, filling forms, and reading rendered content the way a person would. UnleashX runs it for you as a hosted tool, so there's no browser to install and no infrastructure to manage.

Once added to UnleashX, your agent can open a URL, click buttons and links, type into form fields, capture a screenshot, save a page as a PDF, and extract the visible text or full content of a page. It runs against UnleashX's built-in browser, so it works out of the box.

Adding Playwright to UnleashX gives your agent hands and eyes on the live web. Instead of being limited to what it already knows, it can visit a page, interact with it, and turn what it finds into content you can use in a conversation or workflow.

## 2. What you'll need

* An UnleashX account with access to **Data Connectors** and **Workflows**.
* The URL(s) of the pages you want the agent to work with.

<Note>
  There's **nothing to connect**. Playwright is a hosted browser tool built into UnleashX — no account, no API key, and no OAuth. The agent uses UnleashX's own browser.
</Note>

<Warning>
  Playwright reaches **public web pages** as an anonymous visitor. It can't sign in to sites on your behalf, and some sites block automated browsers or challenge them with anti-bot checks (see Troubleshooting).
</Warning>

## 3. Get your credentials

**No credentials needed.** Playwright uses UnleashX's built-in hosted browser — there's no external account to create, no key to paste, and no OAuth flow. There's nothing to authorize.

### What it can do

| Access              | Why UnleashX needs it                          |
| ------------------- | ---------------------------------------------- |
| Open pages          | Navigate the hosted browser to any public URL. |
| Interact with pages | Click elements and fill form fields.           |
| Read content        | Extract visible text and full page content.    |
| Capture pages       | Take screenshots and save pages as PDF.        |

## 4. Connect on UnleashX

Because Playwright is built in, there's nothing to link under **Data Connectors** — you add it directly in the **Workflows** builder (or use it in a conversation) with no setup.

### Use Playwright in a Workflow

Once connected, you can add **Playwright** to any automation from the **Workflows** builder. Its tools appear in the Apps panel, marked with an **MCP** badge.

<Steps>
  <Step title="Add a trigger node">
    Open **Workflows → New Workflow**. On the canvas, click **+ Add Trigger**. In the **Paths** panel, open the **Apps** tab and select **Playwright** — its **Triggers** are listed underneath. Use the search box if you have many connectors.
  </Step>

  <Step title="Add an action node">
    Click the **+** below any node to add a step, then pick **Playwright** again — this time the panel lists its **Actions**.
  </Step>

  <Step title="Configure the step">
    Fill in the fields for the trigger or action you picked. Required fields are marked with a red asterisk (\*).
  </Step>

  <Step title="Add or select your account">
    Under **Selected account**, choose an already-connected account, or click **Add Account** to connect one now.
  </Step>

  <Step title="Save and test">
    Click **Save**. Use **Test** to verify the step, then toggle **Publish** when the workflow is ready.
  </Step>
</Steps>

<Note>
  The steps are the same for every connector. For the full workflow builder guide, see [Using MCP in Workflows](/mcp/workflows).
</Note>

## 5. Available tools

| Tool             | What it does                                     | Changes data? |
| ---------------- | ------------------------------------------------ | ------------- |
| Navigate         | Opens a URL in the hosted browser.               | No            |
| Screenshot       | Captures a screenshot of the current page.       | No            |
| Extract Text     | Extracts the visible text from the current page. | No            |
| Click            | Clicks an element on the page.                   | No            |
| Fill             | Types a value into a form field.                 | No            |
| Get Page Content | Returns the rendered HTML/content of the page.   | No            |
| Page PDF         | Saves the current page as a PDF.                 | No            |
| Get Page Title   | Returns the title of the current page.           | No            |

<Note>
  These tools read and interact with external websites; they don't change your own data. Clicking and filling act on the page in the hosted browser session, not on any UnleashX record.
</Note>

## 6. Example usage

**"Open our competitor's pricing page and tell me their plan tiers and prices."**
→ Runs Navigate to the URL, then Extract Text to read the pricing tiers from the page.

**"Go to this form, fill in my name and email, and take a screenshot of the completed form."**
→ Runs Navigate to the page, Fill for each field, then Screenshot to capture the result.

## 7. Permissions & data access

**UnleashX can:**

* Open public web pages in the hosted browser.
* Click elements, fill form fields, and read page text and content.
* Capture screenshots and save pages as PDF.

**UnleashX cannot:**

* Sign in to sites or access pages behind a login.
* Bypass a site's anti-bot protection, CAPTCHAs, or terms of use.
* Reach pages on private networks it can't route to.

**To disconnect:** Playwright has no stored credentials to revoke. To stop using it, remove the Playwright step from your workflow (or don't add it). There's nothing to disconnect under **Data Connectors**.

## 8. Troubleshooting

| Problem                          | What it means                                        | How to fix it                                                          |
| -------------------------------- | ---------------------------------------------------- | ---------------------------------------------------------------------- |
| **Page timeout**                 | The site was slow or never finished loading.         | Retry, or point the agent at a lighter, more direct URL.               |
| **Selector not found**           | The element to click or fill wasn't on the page.     | Check the element description; the page layout may have changed.       |
| **Blocked / anti-bot page**      | The site detected automation and served a challenge. | The page may be unreachable this way — try a different source.         |
| **JS-heavy page returns little** | Content loads dynamically after the initial render.  | Wait for the page to settle, then re-extract, or capture a screenshot. |
| **Empty content**                | The page rendered nothing extractable.               | Try Get Page Content, or take a screenshot instead of extracting text. |

For general MCP issues, see [/mcp/integrations](/mcp/integrations).

## 9. Frequently asked questions

**Do I need to install or connect anything?**
No. Playwright is a hosted browser built into UnleashX — no account, no key, no OAuth. Just add it in Workflows.

**Can it log in to sites for me?**
No. It browses as an anonymous visitor and can't sign in on your behalf or reach pages behind a login.

**Does UnleashX store the pages it visits?**
No. Pages are loaded live in the hosted browser per request and returned to your agent; UnleashX doesn't keep a copy.

**Is automated browsing allowed?**
Respect each site's terms of service and robots rules. You're responsible for how the agent uses the tool.

## 10. References

* Playwright documentation: [https://playwright.dev/docs/intro](https://playwright.dev/docs/intro)
* Playwright API reference: [https://playwright.dev/docs/api/class-playwright](https://playwright.dev/docs/api/class-playwright)
* Using MCP in Workflows: [https://www.tryunleashx.com](https://www.tryunleashx.com)