> ## Documentation Index
> Fetch the complete documentation index at: https://notte-experiment-visibility-md-links.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Anti-Detection Playbook

> How to avoid CAPTCHAs and bot detection in Notte Sessions

### Resolving Bot Detection Issues

Bot detection is a common challenge for web agents. This guide provides strategies to bypass these detection mechanisms using Notte's stealth configuration options.

Common Bot Detection Challenges:

* Accessing e-commerce sites with anti-bot measures
* Scraping content from news or social media platforms
* Interacting with banking or financial websites
* Accessing sites with geographic restrictions

### Stealth Configuration Strategies

#### 1. Proxy Configuration

Proxies are one of the most effective ways to bypass bot detection. Different proxy configurations can help you appear as legitimate traffic from various locations.

#### Using Default Proxies

Enable Notte's built-in residential proxies for better anonymity:

{/* @sniptest testers/sessions/default_proxy.py */}

```python default_proxy.py theme={null}
from notte_sdk import NotteClient

client = NotteClient()

# Start a session with built-in proxies
with client.Session(proxies=True) as session:
    _ = session.observe(url="https://www.notte.cc/")
```

#### Country-Specific Proxies

For sites with geographic restrictions, use proxies from specific countries:

{/* @sniptest testers/sessions/country_proxy.py */}

```python country_proxy.py theme={null}
from notte_sdk.types import NotteProxy

proxies = NotteProxy.from_country("fr")
```

### 2. Browser Type Selection

Different browsers have varying levels of detection resistance. Experiment with different browser types for your specific use case:

{/* @sniptest testers/guides/browser_type_selection.py */}

```python browser_type_selection.py theme={null}
from typing import Literal

from notte_sdk import NotteClient

client = NotteClient()

# Try different browser types
BrowserType = Literal["chromium", "chrome"]
browsers: list[BrowserType] = ["chromium", "chrome"]
for browser in browsers:
    with client.Session(browser_type=browser, proxies=True, solve_captchas=True) as session:
        result = session.observe(url="https://example.com")
        print(f"Success with {browser}")
```

<Tip>
  `chromium` is the default browser type but is the most easily detected.
</Tip>

### 3. CAPTCHA Solving

Enable automatic CAPTCHA solving for sites that use these challenges:

{/* @sniptest testers/sessions/solve_captchas.py */}

```python solve_captchas.py theme={null}
from notte_sdk import NotteClient

client = NotteClient()
with client.Session(solve_captchas=True, open_viewer=True) as session:
    # Navigate to a page with a CAPTCHA
    agent = client.Agent(session=session, max_steps=5)
    resp = agent.run(
        task=("Try to solve the CAPTCHA using internal tools. If you fail, try to solve it manually."),
        url="https://www.google.com/recaptcha/api2/demo",
    )
```

<Warning>
  Not all CAPTCHA types are supported. Some complex CAPTCHAs may still require manual intervention.
</Warning>

## Complete Stealth Configuration Example

Here's a comprehensive example combining all stealth techniques:

{/* @sniptest testers/sessions/stealth_configuration.py */}

```python stealth_configuration.py theme={null}
from notte_sdk import NotteClient

client = NotteClient()

# Example stealth configuration
# this is just one possible configuration, with an obvious fingerprint
# rotating those values will raise your chances
with client.Session(
    solve_captchas=True,
    proxies="us",
    user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36",
    viewport_width=1920,
    viewport_height=1080,
) as session:
    result = session.observe(url="https://example.com")
    print("Success with fallback configuration")
```

## Troubleshooting Tips

1. **Start Simple**. Begin with basic configurations and gradually add complexity:
   * Try `proxies=True` first
   * Add `solve_captchas=True` if needed
   * Experiment with different `browser_type` values
   * Add custom `user_agent` if still detected
2. **Test Incrementally**. Test each configuration change individually to identify what works:
3. **Monitor for Patterns**. Keep track of which configurations work for different types of sites:
   * E-commerce sites often respond well to residential proxies
   * Social media sites may require specific user agents
   * Banking sites may need country-specific proxies

## Best Practices

1. **Rotate Configurations**: Don't rely on a single configuration: it makes it easier to track you
2. **Monitor Success Rates**: Verify which configurations work best for different site types
3. **Respect Rate Limits**: Implement delays between requests to avoid triggering rate limiting
4. **Keep Configurations Updated**: Bot detection methods evolve, so regularly test and update your configurations