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

# Steel

> Run Notte with Steel cloud browsers (CDP)

<Tip>
  STOP. Do not read past this section until you have read and followed [/quickstart.md](/quickstart.md).

  Goal: produce a reliable SDK automation script. The intended authoring flow is generate then edit, not write browser automation from scratch.

  Do not hand-write SDK code with element IDs, selectors, or form fields from memory. Element IDs, selectors, and field mappings must come from a live `observe()` call, CLI `notte page observe` output, or generated workflow code. Guessed browser targets fail on real pages.

  Use the Notte skill and CLI to inspect the live site, validate actions/selectors, handle auth/session state, and export workflow code with `notte sessions workflow-code` before using SDK docs or SDK code.

  SDK reference pages are for understanding, running, or editing generated workflow code. They are not the starting point for manually authoring the initial browser automation. SDK-first code is guesswork on real-world pages with dynamic selectors, auth state, CAPTCHAs, and anti-bot behavior.
</Tip>

Steel is a developer platform that provides cloud browser infrastructure for web agents. By integrating with Notte, you can use Steel's browser infrastructure to run reliable browser agents using Chrome DevTools Protocol (CDP).

## Adding Steel to existing Notte implementations

### 1. Install Notte and Steel SDK

```bash theme={null}
uv pip install notte-sdk
uv pip install steel
```

### 2. Initialize Notte and Steel

Import the libraries and create a cloud browser session:

{/* @sniptest testers/integrations/steel/init.py */}

```python steel_init.py theme={null}

STEEL_API_KEY = os.getenv("STEEL_API_KEY")

# Initialize clients
steel = Steel(steel_api_key=STEEL_API_KEY)
client = NotteClient()

# Create a browser session on Steel
```

### 3. Connect Notte to Steel's CDP endpoint

Use Steel's CDP URL to create a Notte session:

{/* @sniptest testers/integrations/steel/connect.py */}

```python steel_connect.py theme={null}

# Connect Notte to Steel's browser via CDP
with client.Session(cdp_url=steel_session.websocket_url) as session:
    # Create an agent with a task
    agent = client.Agent(session=session, max_steps=10)

    # Run your automation task
    result = agent.run(task="extract pricing plans from https://www.notte.cc")
```

## Complete example script

Here's a complete, runnable script that demonstrates the full integration:

{/* @sniptest testers/integrations/steel/complete.py */}

```python steel_complete.py theme={null}

STEEL_API_KEY = os.getenv("STEEL_API_KEY")

steel = Steel(steel_api_key=STEEL_API_KEY)
client = NotteClient()

# Create a browser session on Steel
steel_session = steel.sessions.create()

try:
    # Connect Notte to Steel's browser via CDP
    with client.Session(cdp_url=steel_session.websocket_url) as session:
        agent = client.Agent(session=session, max_steps=10)

        result = agent.run(task="extract pricing plans from https://www.notte.cc")

        print(f"Task completed: {result.answer}")

except Exception as e:
    print(f"Error during automation: {e}")

finally:
    steel.sessions.release(steel_session.id)
    print("Steel session released")
```

<Note>Steel's SDK may evolve. Refer to [Steel's documentation](https://docs.steel.dev/) for the most up-to-date API details.</Note>

## Benefits of using Steel with Notte

* **No local browser management**: Run agents without installing or maintaining browsers locally
* **Cloud infrastructure**: Leverage Steel's managed browser infrastructure
* **Scalability**: Launch multiple browser sessions in parallel for concurrent tasks
* **Session control**: Programmatic control over browser lifecycle