Skip to content

armgit / cua

Public Imported from GitHub
0

Scale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.

new usecase

Sarina Liaa2d1f4db9b0
diff --git a/docs/content/docs/example-usecases/meta.json b/docs/content/docs/example-usecases/meta.json
index 2d01953cf5..c7ec3895f9 100644
--- a/docs/content/docs/example-usecases/meta.json
+++ b/docs/content/docs/example-usecases/meta.json
@@ -1,5 +1,5 @@
{
"title": "Cookbook",
"description": "Real-world examples of building with Cua",
- "pages": ["form-filling"]
+ "pages": ["form-filling", "post-event-contact-export"]
}
diff --git a/docs/content/docs/example-usecases/post-event-contact-export.mdx b/docs/content/docs/example-usecases/post-event-contact-export.mdx
new file mode 100644
index 0000000000..fcc6e3f738
--- /dev/null
+++ b/docs/content/docs/example-usecases/post-event-contact-export.mdx
@@ -0,0 +1,1523 @@
+---
+title: Post-Event Contact Export
+description: Run overnight contact extraction from LinkedIn, X, or other social platforms after networking events
+---
+
+import { EditableCodeBlock, EditableValue, S } from '@/components/editable-code-block';
+import { Tab, Tabs } from 'fumadocs-ui/components/tabs';
+
+## Overview
+
+After networking events, you need to export new connections from LinkedIn, X, or other platforms into your CRM. This automation handles it for you.
+
+**The workflow**: Kick off the script after an event and let it run overnight. Wake up to a clean CSV ready for your CRM or email tool.
+
+This example focuses on LinkedIn but works across platforms. It uses [Cua Computer](/computer-sdk/computers) to interact with web interfaces and [Agent Loops](/agent-sdk/agent-loops) to iterate through connections with conversation history.
+
+### Why Cua is Perfect for This
+
+**Cua's VMs save your session data**, bypassing bot detection entirely:
+
+- **Log in once manually** through the VM browser
+- **Session persists** - you appear as a regular user, not a bot
+- **No captchas** - the platform treats automation like normal browsing
+- **No login code** - script doesn't handle authentication
+- **Run overnight** - kick off and forget
+
+Traditional web scraping triggers anti-bot measures immediately. Cua's approach works across all platforms.
+
+## What You Get
+
+The script generates two files with your extracted connections:
+
+**CSV Export** (`linkedin_connections_20250116_143022.csv`):
+
+```csv
+first,last,role,company,met_at,linkedin
+John,Smith,Software Engineer,Acme Corp,Google Devfest Toronto,https://www.linkedin.com/in/johnsmith
+Sarah,Johnson,Product Manager,Tech Inc,Google Devfest Toronto,https://www.linkedin.com/in/sarahjohnson
+Michael,Chen,Data Scientist,StartupXYZ,Google Devfest Toronto,https://www.linkedin.com/in/michaelchen
+Emily,Rodriguez,UX Designer,Design Co,Google Devfest Toronto,https://www.linkedin.com/in/emilyrodriguez
+David,Kim,Engineering Lead,BigTech,Google Devfest Toronto,https://www.linkedin.com/in/davidkim
+...
+```
+
+**Messaging Links** (`linkedin_messaging_links_20250116_143022.txt`):
+
+```
+LinkedIn Messaging Compose Links
+================================================================================
+
+1. https://www.linkedin.com/messaging/compose/?recipient=johnsmith
+2. https://www.linkedin.com/messaging/compose/?recipient=sarahjohnson
+3. https://www.linkedin.com/messaging/compose/?recipient=michaelchen
+4. https://www.linkedin.com/messaging/compose/?recipient=emilyrodriguez
+5. https://www.linkedin.com/messaging/compose/?recipient=davidkim
+...
+```
+
+## Quickstart
+
+Create a `requirements.txt` file with the following dependencies:
+
+```text
+cua-agent
+cua-computer
+python-dotenv>=1.0.0
+```
+
+And install:
+
+```bash
+pip install -r requirements.txt
+```
+
+Create a `.env` file with the following environment variables:
+
+```text
+ANTHROPIC_API_KEY=your-api-key
+CUA_API_KEY=sk_cua-api01...
+CUA_CONTAINER_NAME=m-linux-...
+```
+
+**Important**: Before running the script, manually log into LinkedIn through your VM:
+
+1. Access your VM through the Cua dashboard
+2. Open a browser and navigate to LinkedIn
+3. Log in with your credentials (handle any captchas manually)
+4. Close the browser but leave the VM running
+5. Your session is now saved and ready for automation!
+
+**Configuration**: Customize the script by editing these variables:
+
+```python
+# Where you met these connections (automatically added to CSV)
+MET_AT_REASON = "Google Devfest Toronto"
+
+# Number of contacts to extract (line 134)
+for contact_num in range(1, 21): # Change 21 to extract more/fewer contacts
+```
+
+Select the environment you want to run the code in (_click on the underlined values in the code to edit them directly!_):
+
+<Tabs items={['☁️ Cloud', '🐳 Docker', '🍎 Lume', '🪟 Windows Sandbox']}>
+ <Tab value="☁️ Cloud">
+
+<EditableCodeBlock
+ key="cloud-tab"
+ lang="python"
+ defaultValues={{
+ "sandbox-name": "m-linux-...",
+ "api_key": "sk_cua-api01...",
+ "met-at": "Google Devfest Toronto"
+ }}
+>
+{`import asyncio
+import csv
+import logging
+import os
+import signal
+import traceback
+from datetime import datetime
+
+from agent import ComputerAgent
+from computer import Computer, VMProviderType
+from dotenv import load_dotenv
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(**name**)
+
+# Configuration: Define where you met these connections
+
+MET_AT_REASON = "`}<EditableValue placeholder="met-at" />{`"
+
+def handle_sigint(sig, frame):
+print("\\n\\nExecution interrupted by user. Exiting gracefully...")
+exit(0)
+
+def extract_public_id_from_linkedin_url(linkedin_url):
+"""
+Extract public ID from LinkedIn profile URL.
+Example: https://www.linkedin.com/in/taylor-r-devries/?lipi=... -> taylor-r-devries
+"""
+if not linkedin_url:
+return None
+
+ # Remove query parameters and trailing slashes
+ url = linkedin_url.split('?')[0].rstrip('/')
+
+ # Extract the part after /in/
+ if '/in/' in url:
+ public_id = url.split('/in/')[-1]
+ return public_id
+
+ return None
+
+def extract_contact_from_response(result_output):
+"""
+Extract contact information from agent's response.
+Expects the agent to return data in format:
+FIRST: value
+LAST: value
+ROLE: value
+COMPANY: value
+LINKEDIN: value
+
+ Note: met_at is auto-filled from MET_AT_REASON constant.
+ """
+ contact = {
+ 'first': '',
+ 'last': '',
+ 'role': '',
+ 'company': '',
+ 'met_at': MET_AT_REASON, # Auto-fill from constant
+ 'linkedin': ''
+ }
+
+ # Collect all text from messages for debugging
+ all_text = []
+
+ for item in result_output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ text = content_part.get("text", "")
+ if text:
+ all_text.append(text)
+ # Parse structured output - look for the exact format
+ for line in text.split('\\n'):
+ line = line.strip()
+ # Use case-insensitive matching and handle extra whitespace
+ line_upper = line.upper()
+
+ if line_upper.startswith("FIRST:"):
+ value = line[6:].strip() # Skip "FIRST:" prefix
+ if value and value.upper() != "N/A":
+ contact['first'] = value
+ elif line_upper.startswith("LAST:"):
+ value = line[5:].strip() # Skip "LAST:" prefix
+ if value and value.upper() != "N/A":
+ contact['last'] = value
+ elif line_upper.startswith("ROLE:"):
+ value = line[5:].strip() # Skip "ROLE:" prefix
+ if value and value.upper() != "N/A":
+ contact['role'] = value
+ elif line_upper.startswith("COMPANY:"):
+ value = line[8:].strip() # Skip "COMPANY:" prefix
+ if value and value.upper() != "N/A":
+ contact['company'] = value
+ elif line_upper.startswith("LINKEDIN:"):
+ value = line[9:].strip() # Skip "LINKEDIN:" prefix
+ if value and value.upper() != "N/A":
+ contact['linkedin'] = value
+
+ # Debug logging
+ if not (contact['first'] or contact['last'] or contact['linkedin']):
+ logger.debug(f"Failed to extract. Full text content ({len(all_text)} messages):")
+ for i, text in enumerate(all_text[-3:]): # Show last 3 messages
+ logger.debug(f" Message {i}: {text[:200]}")
+
+ return contact
+
+async def scrape_linkedin_connections():
+"""
+Scrape the first 20 connections from LinkedIn and export to CSV.
+The agent extracts data, and Python handles CSV writing programmatically.
+"""
+
+ # Generate output filename with timestamp
+ timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+ csv_filename = f"linkedin_connections_{timestamp}.csv"
+ csv_path = os.path.join(os.getcwd(), csv_filename)
+
+ # Initialize CSV file with headers
+ with open(csv_path, 'w', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writeheader()
+
+ print(f"\\n🚀 Starting LinkedIn connections scraper")
+ print(f"📁 Output file: {csv_path}")
+ print(f"📍 Met at: {MET_AT_REASON}")
+ print("=" * 80)
+
+ try:
+ async with Computer(
+ os_type="linux",
+ provider_type=VMProviderType.CLOUD,
+ name="`}<EditableValue placeholder="sandbox-name" />{`",
+ api_key="`}<EditableValue placeholder="api_key" />{`",
+ verbosity=logging.INFO,
+ ) as computer:
+
+ agent = ComputerAgent(
+ model="anthropic/claude-sonnet-4-5-20250929",
+ tools=[computer],
+ only_n_most_recent_images=3,
+ verbosity=logging.INFO,
+ trajectory_dir="trajectories",
+ use_prompt_caching=True,
+ max_trajectory_budget=10.0,
+ )
+
+ history = []
+
+ # Task 1: Navigate to LinkedIn connections page
+ navigation_task = (
+ "STEP 1 - NAVIGATE TO LINKEDIN CONNECTIONS PAGE:\\n"
+ "1. Open a web browser (Chrome or Firefox)\\n"
+ "2. Navigate to https://www.linkedin.com/mynetwork/invite-connect/connections/\\n"
+ "3. Wait for the page to fully load (look for the connection list to appear)\\n"
+ "4. If prompted to log in, handle the authentication\\n"
+ "5. Confirm you can see the list of connections displayed on the page\\n"
+ "6. Ready to start extracting contacts one by one"
+ )
+
+ print(f"\\n[Task 1/21] Navigating to LinkedIn connections page...")
+ history.append({"role": "user", "content": navigation_task})
+
+ async for result in agent.run(history, stream=False):
+ history += result.get("output", [])
+ for item in result.get("output", []):
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ print(f"✅ Navigation completed\\n")
+
+ # Tasks 2-21: Extract each of the 20 contacts
+ contacts_extracted = 0
+ linkedin_urls = [] # Track LinkedIn URLs for bonus messaging links
+ previous_contact_name = None # Track the previous contact's name for easy navigation
+
+ for contact_num in range(1, 21):
+ # Build extraction task based on whether this is the first contact or not
+ if contact_num == 1:
+ # First contact - start from the top
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Look at the very first connection at the top of the list\\n"
+ f"2. Click on their name/profile link to open their LinkedIn profile page\\n"
+ f"3. Wait for their profile page to load completely\\n"
+ f"4. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"5. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"6. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"7. Do NOT add any extra text before or after these 5 lines\\n"
+ f"8. Navigate back to the connections list page"
+ )
+ else:
+ # Subsequent contacts - reference the previous contact
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Find the contact named '{previous_contact_name}' in the list\\n"
+ f"2. If you don't see '{previous_contact_name}' on the screen, scroll down slowly until you find them\\n"
+ f"3. Once you find '{previous_contact_name}', look at the contact directly BELOW them\\n"
+ f"4. Click on that contact's name/profile link (the one below '{previous_contact_name}') to open their profile page\\n"
+ f"5. Wait for their profile page to load completely\\n"
+ f"6. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"7. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"8. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"9. Do NOT add any extra text before or after these 5 lines\\n"
+ f"10. Navigate back to the connections list page"
+ )
+
+ print(f"[Task {contact_num + 1}/21] Extracting contact {contact_num}/20...")
+ history.append({"role": "user", "content": extraction_task})
+
+ # Collect all output from the agent
+ all_output = []
+ async for result in agent.run(history, stream=False):
+ output = result.get("output", [])
+ history += output
+ all_output.extend(output)
+
+ # Log agent output at debug level (only shown if verbosity increased)
+ for item in output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ # Now extract contact information from ALL collected output (not just partial results)
+ contact_data = extract_contact_from_response(all_output)
+
+ # Validate we got at least the critical fields (name or LinkedIn URL)
+ has_name = bool(contact_data['first'] and contact_data['last'])
+ has_linkedin = bool(contact_data['linkedin'] and 'linkedin.com' in contact_data['linkedin'])
+
+ # Write to CSV if we got at least name OR linkedin
+ if has_name or has_linkedin:
+ with open(csv_path, 'a', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writerow(contact_data)
+ contacts_extracted += 1
+
+ # Track LinkedIn URL for messaging links
+ if contact_data['linkedin']:
+ linkedin_urls.append(contact_data['linkedin'])
+
+ # Remember this contact's name for the next iteration
+ if has_name:
+ previous_contact_name = f"{contact_data['first']} {contact_data['last']}".strip()
+
+ # Success message with what we got
+ name_str = f"{contact_data['first']} {contact_data['last']}" if has_name else "[No name]"
+ linkedin_str = "✓ LinkedIn" if has_linkedin else "✗ No LinkedIn"
+ role_str = f"({contact_data['role']})" if contact_data['role'] else "(No role)"
+ print(f"✅ Contact {contact_num}/20 saved: {name_str} {role_str} | {linkedin_str}")
+ else:
+ print(f"⚠️ Could not extract valid data for contact {contact_num}")
+ print(f" Got: first='{contact_data['first']}', last='{contact_data['last']}', linkedin='{contact_data['linkedin'][:50] if contact_data['linkedin'] else 'None'}'")
+ print(f" Check the agent's output above to see what was returned")
+ print(f" Total output items: {len(all_output)}")
+
+ # Progress update every 5 contacts
+ if contact_num % 5 == 0:
+ print(f"\\n📈 Progress: {contacts_extracted}/{contact_num} contacts extracted so far...\\n")
+
+ # BONUS: Create messaging compose links file
+ messaging_filename = f"linkedin_messaging_links_{timestamp}.txt"
+ messaging_path = os.path.join(os.getcwd(), messaging_filename)
+
+ with open(messaging_path, 'w', encoding='utf-8') as txtfile:
+ txtfile.write("LinkedIn Messaging Compose Links\\n")
+ txtfile.write("=" * 80 + "\\n\\n")
+
+ for i, linkedin_url in enumerate(linkedin_urls, 1):
+ public_id = extract_public_id_from_linkedin_url(linkedin_url)
+ if public_id:
+ messaging_url = f"https://www.linkedin.com/messaging/compose/?recipient={public_id}"
+ txtfile.write(f"{i}. {messaging_url}\\n")
+ else:
+ txtfile.write(f"{i}. [Could not extract public ID from: {linkedin_url}]\\n")
+
+ print("\\n" + "="*80)
+ print("🎉 All tasks completed!")
+ print(f"📁 CSV file saved to: {csv_path}")
+ print(f"📊 Total contacts extracted: {contacts_extracted}/20")
+ print(f"💬 Bonus: Messaging links saved to: {messaging_path}")
+ print(f"📝 Total messaging links: {len(linkedin_urls)}")
+ print("="*80)
+
+ except Exception as e:
+ print(f"\\n❌ Error during scraping: {e}")
+ traceback.print_exc()
+ raise
+
+def main():
+try:
+load_dotenv()
+
+ if "ANTHROPIC_API_KEY" not in os.environ:
+ raise RuntimeError(
+ "Please set the ANTHROPIC_API_KEY environment variable.\\n"
+ "You can add it to a .env file in the project root."
+ )
+
+ if "CUA_API_KEY" not in os.environ:
+ raise RuntimeError(
+ "Please set the CUA_API_KEY environment variable.\\n"
+ "You can add it to a .env file in the project root."
+ )
+
+ signal.signal(signal.SIGINT, handle_sigint)
+
+ asyncio.run(scrape_linkedin_connections())
+
+ except Exception as e:
+ print(f"\\n❌ Error running automation: {e}")
+ traceback.print_exc()
+
+if **name** == "**main**":
+main()`}
+
+</EditableCodeBlock>
+
+ </Tab>
+ <Tab value="🐳 Docker">
+
+<EditableCodeBlock
+ key="docker-tab"
+ lang="python"
+ defaultValues={{
+ "sandbox-name": "trycua/cua-ubuntu:latest",
+ "met-at": "Google Devfest Toronto"
+ }}
+>
+{`import asyncio
+import csv
+import logging
+import os
+import signal
+import traceback
+from datetime import datetime
+
+from agent import ComputerAgent
+from computer import Computer, VMProviderType
+from dotenv import load_dotenv
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(**name**)
+
+# Configuration: Define where you met these connections
+
+MET_AT_REASON = "`}<EditableValue placeholder="met-at" />{`"
+
+def handle_sigint(sig, frame):
+print("\\n\\nExecution interrupted by user. Exiting gracefully...")
+exit(0)
+
+def extract_public_id_from_linkedin_url(linkedin_url):
+"""
+Extract public ID from LinkedIn profile URL.
+Example: https://www.linkedin.com/in/taylor-r-devries/?lipi=... -> taylor-r-devries
+"""
+if not linkedin_url:
+return None
+
+ # Remove query parameters and trailing slashes
+ url = linkedin_url.split('?')[0].rstrip('/')
+
+ # Extract the part after /in/
+ if '/in/' in url:
+ public_id = url.split('/in/')[-1]
+ return public_id
+
+ return None
+
+def extract_contact_from_response(result_output):
+"""
+Extract contact information from agent's response.
+Expects the agent to return data in format:
+FIRST: value
+LAST: value
+ROLE: value
+COMPANY: value
+LINKEDIN: value
+
+ Note: met_at is auto-filled from MET_AT_REASON constant.
+ """
+ contact = {
+ 'first': '',
+ 'last': '',
+ 'role': '',
+ 'company': '',
+ 'met_at': MET_AT_REASON, # Auto-fill from constant
+ 'linkedin': ''
+ }
+
+ # Collect all text from messages for debugging
+ all_text = []
+
+ for item in result_output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ text = content_part.get("text", "")
+ if text:
+ all_text.append(text)
+ # Parse structured output - look for the exact format
+ for line in text.split('\\n'):
+ line = line.strip()
+ # Use case-insensitive matching and handle extra whitespace
+ line_upper = line.upper()
+
+ if line_upper.startswith("FIRST:"):
+ value = line[6:].strip() # Skip "FIRST:" prefix
+ if value and value.upper() != "N/A":
+ contact['first'] = value
+ elif line_upper.startswith("LAST:"):
+ value = line[5:].strip() # Skip "LAST:" prefix
+ if value and value.upper() != "N/A":
+ contact['last'] = value
+ elif line_upper.startswith("ROLE:"):
+ value = line[5:].strip() # Skip "ROLE:" prefix
+ if value and value.upper() != "N/A":
+ contact['role'] = value
+ elif line_upper.startswith("COMPANY:"):
+ value = line[8:].strip() # Skip "COMPANY:" prefix
+ if value and value.upper() != "N/A":
+ contact['company'] = value
+ elif line_upper.startswith("LINKEDIN:"):
+ value = line[9:].strip() # Skip "LINKEDIN:" prefix
+ if value and value.upper() != "N/A":
+ contact['linkedin'] = value
+
+ # Debug logging
+ if not (contact['first'] or contact['last'] or contact['linkedin']):
+ logger.debug(f"Failed to extract. Full text content ({len(all_text)} messages):")
+ for i, text in enumerate(all_text[-3:]): # Show last 3 messages
+ logger.debug(f" Message {i}: {text[:200]}")
+
+ return contact
+
+async def scrape_linkedin_connections():
+"""
+Scrape the first 20 connections from LinkedIn and export to CSV.
+The agent extracts data, and Python handles CSV writing programmatically.
+"""
+
+ # Generate output filename with timestamp
+ timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+ csv_filename = f"linkedin_connections_{timestamp}.csv"
+ csv_path = os.path.join(os.getcwd(), csv_filename)
+
+ # Initialize CSV file with headers
+ with open(csv_path, 'w', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writeheader()
+
+ print(f"\\n🚀 Starting LinkedIn connections scraper")
+ print(f"📁 Output file: {csv_path}")
+ print(f"📍 Met at: {MET_AT_REASON}")
+ print("=" * 80)
+
+ try:
+ async with Computer(
+ os_type="linux",
+ provider_type=VMProviderType.DOCKER,
+ name="`}<EditableValue placeholder="sandbox-name" />{`",
+ verbosity=logging.INFO,
+ ) as computer:
+
+ agent = ComputerAgent(
+ model="anthropic/claude-sonnet-4-5-20250929",
+ tools=[computer],
+ only_n_most_recent_images=3,
+ verbosity=logging.INFO,
+ trajectory_dir="trajectories",
+ use_prompt_caching=True,
+ max_trajectory_budget=10.0,
+ )
+
+ history = []
+
+ # Task 1: Navigate to LinkedIn connections page
+ navigation_task = (
+ "STEP 1 - NAVIGATE TO LINKEDIN CONNECTIONS PAGE:\\n"
+ "1. Open a web browser (Chrome or Firefox)\\n"
+ "2. Navigate to https://www.linkedin.com/mynetwork/invite-connect/connections/\\n"
+ "3. Wait for the page to fully load (look for the connection list to appear)\\n"
+ "4. If prompted to log in, handle the authentication\\n"
+ "5. Confirm you can see the list of connections displayed on the page\\n"
+ "6. Ready to start extracting contacts one by one"
+ )
+
+ print(f"\\n[Task 1/21] Navigating to LinkedIn connections page...")
+ history.append({"role": "user", "content": navigation_task})
+
+ async for result in agent.run(history, stream=False):
+ history += result.get("output", [])
+ for item in result.get("output", []):
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ print(f"✅ Navigation completed\\n")
+
+ # Tasks 2-21: Extract each of the 20 contacts
+ contacts_extracted = 0
+ linkedin_urls = [] # Track LinkedIn URLs for bonus messaging links
+ previous_contact_name = None # Track the previous contact's name for easy navigation
+
+ for contact_num in range(1, 21):
+ # Build extraction task based on whether this is the first contact or not
+ if contact_num == 1:
+ # First contact - start from the top
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Look at the very first connection at the top of the list\\n"
+ f"2. Click on their name/profile link to open their LinkedIn profile page\\n"
+ f"3. Wait for their profile page to load completely\\n"
+ f"4. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"5. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"6. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"7. Do NOT add any extra text before or after these 5 lines\\n"
+ f"8. Navigate back to the connections list page"
+ )
+ else:
+ # Subsequent contacts - reference the previous contact
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Find the contact named '{previous_contact_name}' in the list\\n"
+ f"2. If you don't see '{previous_contact_name}' on the screen, scroll down slowly until you find them\\n"
+ f"3. Once you find '{previous_contact_name}', look at the contact directly BELOW them\\n"
+ f"4. Click on that contact's name/profile link (the one below '{previous_contact_name}') to open their profile page\\n"
+ f"5. Wait for their profile page to load completely\\n"
+ f"6. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"7. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"8. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"9. Do NOT add any extra text before or after these 5 lines\\n"
+ f"10. Navigate back to the connections list page"
+ )
+
+ print(f"[Task {contact_num + 1}/21] Extracting contact {contact_num}/20...")
+ history.append({"role": "user", "content": extraction_task})
+
+ # Collect all output from the agent
+ all_output = []
+ async for result in agent.run(history, stream=False):
+ output = result.get("output", [])
+ history += output
+ all_output.extend(output)
+
+ # Log agent output at debug level (only shown if verbosity increased)
+ for item in output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ # Now extract contact information from ALL collected output (not just partial results)
+ contact_data = extract_contact_from_response(all_output)
+
+ # Validate we got at least the critical fields (name or LinkedIn URL)
+ has_name = bool(contact_data['first'] and contact_data['last'])
+ has_linkedin = bool(contact_data['linkedin'] and 'linkedin.com' in contact_data['linkedin'])
+
+ # Write to CSV if we got at least name OR linkedin
+ if has_name or has_linkedin:
+ with open(csv_path, 'a', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writerow(contact_data)
+ contacts_extracted += 1
+
+ # Track LinkedIn URL for messaging links
+ if contact_data['linkedin']:
+ linkedin_urls.append(contact_data['linkedin'])
+
+ # Remember this contact's name for the next iteration
+ if has_name:
+ previous_contact_name = f"{contact_data['first']} {contact_data['last']}".strip()
+
+ # Success message with what we got
+ name_str = f"{contact_data['first']} {contact_data['last']}" if has_name else "[No name]"
+ linkedin_str = "✓ LinkedIn" if has_linkedin else "✗ No LinkedIn"
+ role_str = f"({contact_data['role']})" if contact_data['role'] else "(No role)"
+ print(f"✅ Contact {contact_num}/20 saved: {name_str} {role_str} | {linkedin_str}")
+ else:
+ print(f"⚠️ Could not extract valid data for contact {contact_num}")
+ print(f" Got: first='{contact_data['first']}', last='{contact_data['last']}', linkedin='{contact_data['linkedin'][:50] if contact_data['linkedin'] else 'None'}'")
+ print(f" Check the agent's output above to see what was returned")
+ print(f" Total output items: {len(all_output)}")
+
+ # Progress update every 5 contacts
+ if contact_num % 5 == 0:
+ print(f"\\n📈 Progress: {contacts_extracted}/{contact_num} contacts extracted so far...\\n")
+
+ # BONUS: Create messaging compose links file
+ messaging_filename = f"linkedin_messaging_links_{timestamp}.txt"
+ messaging_path = os.path.join(os.getcwd(), messaging_filename)
+
+ with open(messaging_path, 'w', encoding='utf-8') as txtfile:
+ txtfile.write("LinkedIn Messaging Compose Links\\n")
+ txtfile.write("=" * 80 + "\\n\\n")
+
+ for i, linkedin_url in enumerate(linkedin_urls, 1):
+ public_id = extract_public_id_from_linkedin_url(linkedin_url)
+ if public_id:
+ messaging_url = f"https://www.linkedin.com/messaging/compose/?recipient={public_id}"
+ txtfile.write(f"{i}. {messaging_url}\\n")
+ else:
+ txtfile.write(f"{i}. [Could not extract public ID from: {linkedin_url}]\\n")
+
+ print("\\n" + "="*80)
+ print("🎉 All tasks completed!")
+ print(f"📁 CSV file saved to: {csv_path}")
+ print(f"📊 Total contacts extracted: {contacts_extracted}/20")
+ print(f"💬 Bonus: Messaging links saved to: {messaging_path}")
+ print(f"📝 Total messaging links: {len(linkedin_urls)}")
+ print("="*80)
+
+ except Exception as e:
+ print(f"\\n❌ Error during scraping: {e}")
+ traceback.print_exc()
+ raise
+
+def main():
+try:
+load_dotenv()
+
+ if "ANTHROPIC_API_KEY" not in os.environ:
+ raise RuntimeError(
+ "Please set the ANTHROPIC_API_KEY environment variable.\\n"
+ "You can add it to a .env file in the project root."
+ )
+
+ signal.signal(signal.SIGINT, handle_sigint)
+
+ asyncio.run(scrape_linkedin_connections())
+
+ except Exception as e:
+ print(f"\\n❌ Error running automation: {e}")
+ traceback.print_exc()
+
+if **name** == "**main**":
+main()`}
+
+</EditableCodeBlock>
+
+ </Tab>
+ <Tab value="🍎 Lume">
+
+<EditableCodeBlock
+ key="lume-tab"
+ lang="python"
+ defaultValues={{
+ "sandbox-name": "macos-sequoia-cua:latest",
+ "met-at": "Google Devfest Toronto"
+ }}
+>
+{`import asyncio
+import csv
+import logging
+import os
+import signal
+import traceback
+from datetime import datetime
+
+from agent import ComputerAgent
+from computer import Computer, VMProviderType
+from dotenv import load_dotenv
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(**name**)
+
+# Configuration: Define where you met these connections
+
+MET_AT_REASON = "`}<EditableValue placeholder="met-at" />{`"
+
+def handle_sigint(sig, frame):
+print("\\n\\nExecution interrupted by user. Exiting gracefully...")
+exit(0)
+
+def extract_public_id_from_linkedin_url(linkedin_url):
+"""
+Extract public ID from LinkedIn profile URL.
+Example: https://www.linkedin.com/in/taylor-r-devries/?lipi=... -> taylor-r-devries
+"""
+if not linkedin_url:
+return None
+
+ # Remove query parameters and trailing slashes
+ url = linkedin_url.split('?')[0].rstrip('/')
+
+ # Extract the part after /in/
+ if '/in/' in url:
+ public_id = url.split('/in/')[-1]
+ return public_id
+
+ return None
+
+def extract_contact_from_response(result_output):
+"""
+Extract contact information from agent's response.
+Expects the agent to return data in format:
+FIRST: value
+LAST: value
+ROLE: value
+COMPANY: value
+LINKEDIN: value
+
+ Note: met_at is auto-filled from MET_AT_REASON constant.
+ """
+ contact = {
+ 'first': '',
+ 'last': '',
+ 'role': '',
+ 'company': '',
+ 'met_at': MET_AT_REASON, # Auto-fill from constant
+ 'linkedin': ''
+ }
+
+ # Collect all text from messages for debugging
+ all_text = []
+
+ for item in result_output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ text = content_part.get("text", "")
+ if text:
+ all_text.append(text)
+ # Parse structured output - look for the exact format
+ for line in text.split('\\n'):
+ line = line.strip()
+ # Use case-insensitive matching and handle extra whitespace
+ line_upper = line.upper()
+
+ if line_upper.startswith("FIRST:"):
+ value = line[6:].strip() # Skip "FIRST:" prefix
+ if value and value.upper() != "N/A":
+ contact['first'] = value
+ elif line_upper.startswith("LAST:"):
+ value = line[5:].strip() # Skip "LAST:" prefix
+ if value and value.upper() != "N/A":
+ contact['last'] = value
+ elif line_upper.startswith("ROLE:"):
+ value = line[5:].strip() # Skip "ROLE:" prefix
+ if value and value.upper() != "N/A":
+ contact['role'] = value
+ elif line_upper.startswith("COMPANY:"):
+ value = line[8:].strip() # Skip "COMPANY:" prefix
+ if value and value.upper() != "N/A":
+ contact['company'] = value
+ elif line_upper.startswith("LINKEDIN:"):
+ value = line[9:].strip() # Skip "LINKEDIN:" prefix
+ if value and value.upper() != "N/A":
+ contact['linkedin'] = value
+
+ # Debug logging
+ if not (contact['first'] or contact['last'] or contact['linkedin']):
+ logger.debug(f"Failed to extract. Full text content ({len(all_text)} messages):")
+ for i, text in enumerate(all_text[-3:]): # Show last 3 messages
+ logger.debug(f" Message {i}: {text[:200]}")
+
+ return contact
+
+async def scrape_linkedin_connections():
+"""
+Scrape the first 20 connections from LinkedIn and export to CSV.
+The agent extracts data, and Python handles CSV writing programmatically.
+"""
+
+ # Generate output filename with timestamp
+ timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+ csv_filename = f"linkedin_connections_{timestamp}.csv"
+ csv_path = os.path.join(os.getcwd(), csv_filename)
+
+ # Initialize CSV file with headers
+ with open(csv_path, 'w', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writeheader()
+
+ print(f"\\n🚀 Starting LinkedIn connections scraper")
+ print(f"📁 Output file: {csv_path}")
+ print(f"📍 Met at: {MET_AT_REASON}")
+ print("=" * 80)
+
+ try:
+ async with Computer(
+ os_type="macos",
+ provider_type=VMProviderType.LUME,
+ name="`}<EditableValue placeholder="sandbox-name" />{`",
+ verbosity=logging.INFO,
+ ) as computer:
+
+ agent = ComputerAgent(
+ model="anthropic/claude-sonnet-4-5-20250929",
+ tools=[computer],
+ only_n_most_recent_images=3,
+ verbosity=logging.INFO,
+ trajectory_dir="trajectories",
+ use_prompt_caching=True,
+ max_trajectory_budget=10.0,
+ )
+
+ history = []
+
+ # Task 1: Navigate to LinkedIn connections page
+ navigation_task = (
+ "STEP 1 - NAVIGATE TO LINKEDIN CONNECTIONS PAGE:\\n"
+ "1. Open a web browser (Chrome or Firefox)\\n"
+ "2. Navigate to https://www.linkedin.com/mynetwork/invite-connect/connections/\\n"
+ "3. Wait for the page to fully load (look for the connection list to appear)\\n"
+ "4. If prompted to log in, handle the authentication\\n"
+ "5. Confirm you can see the list of connections displayed on the page\\n"
+ "6. Ready to start extracting contacts one by one"
+ )
+
+ print(f"\\n[Task 1/21] Navigating to LinkedIn connections page...")
+ history.append({"role": "user", "content": navigation_task})
+
+ async for result in agent.run(history, stream=False):
+ history += result.get("output", [])
+ for item in result.get("output", []):
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ print(f"✅ Navigation completed\\n")
+
+ # Tasks 2-21: Extract each of the 20 contacts
+ contacts_extracted = 0
+ linkedin_urls = [] # Track LinkedIn URLs for bonus messaging links
+ previous_contact_name = None # Track the previous contact's name for easy navigation
+
+ for contact_num in range(1, 21):
+ # Build extraction task based on whether this is the first contact or not
+ if contact_num == 1:
+ # First contact - start from the top
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Look at the very first connection at the top of the list\\n"
+ f"2. Click on their name/profile link to open their LinkedIn profile page\\n"
+ f"3. Wait for their profile page to load completely\\n"
+ f"4. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"5. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"6. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"7. Do NOT add any extra text before or after these 5 lines\\n"
+ f"8. Navigate back to the connections list page"
+ )
+ else:
+ # Subsequent contacts - reference the previous contact
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Find the contact named '{previous_contact_name}' in the list\\n"
+ f"2. If you don't see '{previous_contact_name}' on the screen, scroll down slowly until you find them\\n"
+ f"3. Once you find '{previous_contact_name}', look at the contact directly BELOW them\\n"
+ f"4. Click on that contact's name/profile link (the one below '{previous_contact_name}') to open their profile page\\n"
+ f"5. Wait for their profile page to load completely\\n"
+ f"6. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"7. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"8. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"9. Do NOT add any extra text before or after these 5 lines\\n"
+ f"10. Navigate back to the connections list page"
+ )
+
+ print(f"[Task {contact_num + 1}/21] Extracting contact {contact_num}/20...")
+ history.append({"role": "user", "content": extraction_task})
+
+ # Collect all output from the agent
+ all_output = []
+ async for result in agent.run(history, stream=False):
+ output = result.get("output", [])
+ history += output
+ all_output.extend(output)
+
+ # Log agent output at debug level (only shown if verbosity increased)
+ for item in output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ # Now extract contact information from ALL collected output (not just partial results)
+ contact_data = extract_contact_from_response(all_output)
+
+ # Validate we got at least the critical fields (name or LinkedIn URL)
+ has_name = bool(contact_data['first'] and contact_data['last'])
+ has_linkedin = bool(contact_data['linkedin'] and 'linkedin.com' in contact_data['linkedin'])
+
+ # Write to CSV if we got at least name OR linkedin
+ if has_name or has_linkedin:
+ with open(csv_path, 'a', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writerow(contact_data)
+ contacts_extracted += 1
+
+ # Track LinkedIn URL for messaging links
+ if contact_data['linkedin']:
+ linkedin_urls.append(contact_data['linkedin'])
+
+ # Remember this contact's name for the next iteration
+ if has_name:
+ previous_contact_name = f"{contact_data['first']} {contact_data['last']}".strip()
+
+ # Success message with what we got
+ name_str = f"{contact_data['first']} {contact_data['last']}" if has_name else "[No name]"
+ linkedin_str = "✓ LinkedIn" if has_linkedin else "✗ No LinkedIn"
+ role_str = f"({contact_data['role']})" if contact_data['role'] else "(No role)"
+ print(f"✅ Contact {contact_num}/20 saved: {name_str} {role_str} | {linkedin_str}")
+ else:
+ print(f"⚠️ Could not extract valid data for contact {contact_num}")
+ print(f" Got: first='{contact_data['first']}', last='{contact_data['last']}', linkedin='{contact_data['linkedin'][:50] if contact_data['linkedin'] else 'None'}'")
+ print(f" Check the agent's output above to see what was returned")
+ print(f" Total output items: {len(all_output)}")
+
+ # Progress update every 5 contacts
+ if contact_num % 5 == 0:
+ print(f"\\n📈 Progress: {contacts_extracted}/{contact_num} contacts extracted so far...\\n")
+
+ # BONUS: Create messaging compose links file
+ messaging_filename = f"linkedin_messaging_links_{timestamp}.txt"
+ messaging_path = os.path.join(os.getcwd(), messaging_filename)
+
+ with open(messaging_path, 'w', encoding='utf-8') as txtfile:
+ txtfile.write("LinkedIn Messaging Compose Links\\n")
+ txtfile.write("=" * 80 + "\\n\\n")
+
+ for i, linkedin_url in enumerate(linkedin_urls, 1):
+ public_id = extract_public_id_from_linkedin_url(linkedin_url)
+ if public_id:
+ messaging_url = f"https://www.linkedin.com/messaging/compose/?recipient={public_id}"
+ txtfile.write(f"{i}. {messaging_url}\\n")
+ else:
+ txtfile.write(f"{i}. [Could not extract public ID from: {linkedin_url}]\\n")
+
+ print("\\n" + "="*80)
+ print("🎉 All tasks completed!")
+ print(f"📁 CSV file saved to: {csv_path}")
+ print(f"📊 Total contacts extracted: {contacts_extracted}/20")
+ print(f"💬 Bonus: Messaging links saved to: {messaging_path}")
+ print(f"📝 Total messaging links: {len(linkedin_urls)}")
+ print("="*80)
+
+ except Exception as e:
+ print(f"\\n❌ Error during scraping: {e}")
+ traceback.print_exc()
+ raise
+
+def main():
+try:
+load_dotenv()
+
+ if "ANTHROPIC_API_KEY" not in os.environ:
+ raise RuntimeError(
+ "Please set the ANTHROPIC_API_KEY environment variable.\\n"
+ "You can add it to a .env file in the project root."
+ )
+
+ signal.signal(signal.SIGINT, handle_sigint)
+
+ asyncio.run(scrape_linkedin_connections())
+
+ except Exception as e:
+ print(f"\\n❌ Error running automation: {e}")
+ traceback.print_exc()
+
+if **name** == "**main**":
+main()`}
+
+</EditableCodeBlock>
+
+ </Tab>
+ <Tab value="🪟 Windows Sandbox">
+
+<EditableCodeBlock
+ key="windows-tab"
+ lang="python"
+ defaultValues={{
+ "met-at": "Google Devfest Toronto"
+ }}
+>
+{`import asyncio
+import csv
+import logging
+import os
+import signal
+import traceback
+from datetime import datetime
+
+from agent import ComputerAgent
+from computer import Computer, VMProviderType
+from dotenv import load_dotenv
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(**name**)
+
+# Configuration: Define where you met these connections
+
+MET_AT_REASON = "`}<EditableValue placeholder="met-at" />{`"
+
+def handle_sigint(sig, frame):
+print("\\n\\nExecution interrupted by user. Exiting gracefully...")
+exit(0)
+
+def extract_public_id_from_linkedin_url(linkedin_url):
+"""
+Extract public ID from LinkedIn profile URL.
+Example: https://www.linkedin.com/in/taylor-r-devries/?lipi=... -> taylor-r-devries
+"""
+if not linkedin_url:
+return None
+
+ # Remove query parameters and trailing slashes
+ url = linkedin_url.split('?')[0].rstrip('/')
+
+ # Extract the part after /in/
+ if '/in/' in url:
+ public_id = url.split('/in/')[-1]
+ return public_id
+
+ return None
+
+def extract_contact_from_response(result_output):
+"""
+Extract contact information from agent's response.
+Expects the agent to return data in format:
+FIRST: value
+LAST: value
+ROLE: value
+COMPANY: value
+LINKEDIN: value
+
+ Note: met_at is auto-filled from MET_AT_REASON constant.
+ """
+ contact = {
+ 'first': '',
+ 'last': '',
+ 'role': '',
+ 'company': '',
+ 'met_at': MET_AT_REASON, # Auto-fill from constant
+ 'linkedin': ''
+ }
+
+ # Collect all text from messages for debugging
+ all_text = []
+
+ for item in result_output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ text = content_part.get("text", "")
+ if text:
+ all_text.append(text)
+ # Parse structured output - look for the exact format
+ for line in text.split('\\n'):
+ line = line.strip()
+ # Use case-insensitive matching and handle extra whitespace
+ line_upper = line.upper()
+
+ if line_upper.startswith("FIRST:"):
+ value = line[6:].strip() # Skip "FIRST:" prefix
+ if value and value.upper() != "N/A":
+ contact['first'] = value
+ elif line_upper.startswith("LAST:"):
+ value = line[5:].strip() # Skip "LAST:" prefix
+ if value and value.upper() != "N/A":
+ contact['last'] = value
+ elif line_upper.startswith("ROLE:"):
+ value = line[5:].strip() # Skip "ROLE:" prefix
+ if value and value.upper() != "N/A":
+ contact['role'] = value
+ elif line_upper.startswith("COMPANY:"):
+ value = line[8:].strip() # Skip "COMPANY:" prefix
+ if value and value.upper() != "N/A":
+ contact['company'] = value
+ elif line_upper.startswith("LINKEDIN:"):
+ value = line[9:].strip() # Skip "LINKEDIN:" prefix
+ if value and value.upper() != "N/A":
+ contact['linkedin'] = value
+
+ # Debug logging
+ if not (contact['first'] or contact['last'] or contact['linkedin']):
+ logger.debug(f"Failed to extract. Full text content ({len(all_text)} messages):")
+ for i, text in enumerate(all_text[-3:]): # Show last 3 messages
+ logger.debug(f" Message {i}: {text[:200]}")
+
+ return contact
+
+async def scrape_linkedin_connections():
+"""
+Scrape the first 20 connections from LinkedIn and export to CSV.
+The agent extracts data, and Python handles CSV writing programmatically.
+"""
+
+ # Generate output filename with timestamp
+ timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+ csv_filename = f"linkedin_connections_{timestamp}.csv"
+ csv_path = os.path.join(os.getcwd(), csv_filename)
+
+ # Initialize CSV file with headers
+ with open(csv_path, 'w', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writeheader()
+
+ print(f"\\n🚀 Starting LinkedIn connections scraper")
+ print(f"📁 Output file: {csv_path}")
+ print(f"📍 Met at: {MET_AT_REASON}")
+ print("=" * 80)
+
+ try:
+ async with Computer(
+ os_type="windows",
+ provider_type=VMProviderType.WINDOWS_SANDBOX,
+ verbosity=logging.INFO,
+ ) as computer:
+
+ agent = ComputerAgent(
+ model="anthropic/claude-sonnet-4-5-20250929",
+ tools=[computer],
+ only_n_most_recent_images=3,
+ verbosity=logging.INFO,
+ trajectory_dir="trajectories",
+ use_prompt_caching=True,
+ max_trajectory_budget=10.0,
+ )
+
+ history = []
+
+ # Task 1: Navigate to LinkedIn connections page
+ navigation_task = (
+ "STEP 1 - NAVIGATE TO LINKEDIN CONNECTIONS PAGE:\\n"
+ "1. Open a web browser (Chrome or Firefox)\\n"
+ "2. Navigate to https://www.linkedin.com/mynetwork/invite-connect/connections/\\n"
+ "3. Wait for the page to fully load (look for the connection list to appear)\\n"
+ "4. If prompted to log in, handle the authentication\\n"
+ "5. Confirm you can see the list of connections displayed on the page\\n"
+ "6. Ready to start extracting contacts one by one"
+ )
+
+ print(f"\\n[Task 1/21] Navigating to LinkedIn connections page...")
+ history.append({"role": "user", "content": navigation_task})
+
+ async for result in agent.run(history, stream=False):
+ history += result.get("output", [])
+ for item in result.get("output", []):
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ print(f"✅ Navigation completed\\n")
+
+ # Tasks 2-21: Extract each of the 20 contacts
+ contacts_extracted = 0
+ linkedin_urls = [] # Track LinkedIn URLs for bonus messaging links
+ previous_contact_name = None # Track the previous contact's name for easy navigation
+
+ for contact_num in range(1, 21):
+ # Build extraction task based on whether this is the first contact or not
+ if contact_num == 1:
+ # First contact - start from the top
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Look at the very first connection at the top of the list\\n"
+ f"2. Click on their name/profile link to open their LinkedIn profile page\\n"
+ f"3. Wait for their profile page to load completely\\n"
+ f"4. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"5. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"6. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"7. Do NOT add any extra text before or after these 5 lines\\n"
+ f"8. Navigate back to the connections list page"
+ )
+ else:
+ # Subsequent contacts - reference the previous contact
+ extraction_task = (
+ f"STEP {contact_num + 1} - EXTRACT CONTACT {contact_num} OF 20:\\n"
+ f"1. Find the contact named '{previous_contact_name}' in the list\\n"
+ f"2. If you don't see '{previous_contact_name}' on the screen, scroll down slowly until you find them\\n"
+ f"3. Once you find '{previous_contact_name}', look at the contact directly BELOW them\\n"
+ f"4. Click on that contact's name/profile link (the one below '{previous_contact_name}') to open their profile page\\n"
+ f"5. Wait for their profile page to load completely\\n"
+ f"6. Extract the following information from their profile:\\n"
+ f" - First name: Extract from their display name at the top (just the first name)\\n"
+ f" - Last name: Extract from their display name at the top (just the last name)\\n"
+ f" - Current role/title: Extract from the HEADLINE directly under their name (e.g., 'Software Engineer')\\n"
+ f" - Company name: Extract from the HEADLINE (typically after 'at' or '@', e.g., 'Software Engineer at Google' → 'Google')\\n"
+ f" - LinkedIn profile URL: Copy the FULL URL from the browser address bar (must start with https://www.linkedin.com/in/)\\n"
+ f"7. CRITICAL: You MUST return ALL 5 fields in this EXACT format with each field on its own line:\\n"
+ f"FIRST: [first name]\\n"
+ f"LAST: [last name]\\n"
+ f"ROLE: [role/title from headline]\\n"
+ f"COMPANY: [company from headline]\\n"
+ f"LINKEDIN: [full profile URL]\\n"
+ f"\\n"
+ f"8. If any field is not available, write 'N/A' instead of leaving it blank\\n"
+ f"9. Do NOT add any extra text before or after these 5 lines\\n"
+ f"10. Navigate back to the connections list page"
+ )
+
+ print(f"[Task {contact_num + 1}/21] Extracting contact {contact_num}/20...")
+ history.append({"role": "user", "content": extraction_task})
+
+ # Collect all output from the agent
+ all_output = []
+ async for result in agent.run(history, stream=False):
+ output = result.get("output", [])
+ history += output
+ all_output.extend(output)
+
+ # Log agent output at debug level (only shown if verbosity increased)
+ for item in output:
+ if item.get("type") == "message":
+ content = item.get("content", [])
+ for content_part in content:
+ if content_part.get("text"):
+ logger.debug(f"Agent: {content_part.get('text')}")
+
+ # Now extract contact information from ALL collected output (not just partial results)
+ contact_data = extract_contact_from_response(all_output)
+
+ # Validate we got at least the critical fields (name or LinkedIn URL)
+ has_name = bool(contact_data['first'] and contact_data['last'])
+ has_linkedin = bool(contact_data['linkedin'] and 'linkedin.com' in contact_data['linkedin'])
+
+ # Write to CSV if we got at least name OR linkedin
+ if has_name or has_linkedin:
+ with open(csv_path, 'a', newline='', encoding='utf-8') as csvfile:
+ writer = csv.DictWriter(csvfile, fieldnames=['first', 'last', 'role', 'company', 'met_at', 'linkedin'])
+ writer.writerow(contact_data)
+ contacts_extracted += 1
+
+ # Track LinkedIn URL for messaging links
+ if contact_data['linkedin']:
+ linkedin_urls.append(contact_data['linkedin'])
+
+ # Remember this contact's name for the next iteration
+ if has_name:
+ previous_contact_name = f"{contact_data['first']} {contact_data['last']}".strip()
+
+ # Success message with what we got
+ name_str = f"{contact_data['first']} {contact_data['last']}" if has_name else "[No name]"
+ linkedin_str = "✓ LinkedIn" if has_linkedin else "✗ No LinkedIn"
+ role_str = f"({contact_data['role']})" if contact_data['role'] else "(No role)"
+ print(f"✅ Contact {contact_num}/20 saved: {name_str} {role_str} | {linkedin_str}")
+ else:
+ print(f"⚠️ Could not extract valid data for contact {contact_num}")
+ print(f" Got: first='{contact_data['first']}', last='{contact_data['last']}', linkedin='{contact_data['linkedin'][:50] if contact_data['linkedin'] else 'None'}'")
+ print(f" Check the agent's output above to see what was returned")
+ print(f" Total output items: {len(all_output)}")
+
+ # Progress update every 5 contacts
+ if contact_num % 5 == 0:
+ print(f"\\n📈 Progress: {contacts_extracted}/{contact_num} contacts extracted so far...\\n")
+
+ # BONUS: Create messaging compose links file
+ messaging_filename = f"linkedin_messaging_links_{timestamp}.txt"
+ messaging_path = os.path.join(os.getcwd(), messaging_filename)
+
+ with open(messaging_path, 'w', encoding='utf-8') as txtfile:
+ txtfile.write("LinkedIn Messaging Compose Links\\n")
+ txtfile.write("=" * 80 + "\\n\\n")
+
+ for i, linkedin_url in enumerate(linkedin_urls, 1):
+ public_id = extract_public_id_from_linkedin_url(linkedin_url)
+ if public_id:
+ messaging_url = f"https://www.linkedin.com/messaging/compose/?recipient={public_id}"
+ txtfile.write(f"{i}. {messaging_url}\\n")
+ else:
+ txtfile.write(f"{i}. [Could not extract public ID from: {linkedin_url}]\\n")
+
+ print("\\n" + "="*80)
+ print("🎉 All tasks completed!")
+ print(f"📁 CSV file saved to: {csv_path}")
+ print(f"📊 Total contacts extracted: {contacts_extracted}/20")
+ print(f"💬 Bonus: Messaging links saved to: {messaging_path}")
+ print(f"📝 Total messaging links: {len(linkedin_urls)}")
+ print("="*80)
+
+ except Exception as e:
+ print(f"\\n❌ Error during scraping: {e}")
+ traceback.print_exc()
+ raise
+
+def main():
+try:
+load_dotenv()
+
+ if "ANTHROPIC_API_KEY" not in os.environ:
+ raise RuntimeError(
+ "Please set the ANTHROPIC_API_KEY environment variable.\\n"
+ "You can add it to a .env file in the project root."
+ )
+
+ signal.signal(signal.SIGINT, handle_sigint)
+
+ asyncio.run(scrape_linkedin_connections())
+
+ except Exception as e:
+ print(f"\\n❌ Error running automation: {e}")
+ traceback.print_exc()
+
+if **name** == "**main**":
+main()`}
+
+</EditableCodeBlock>
+
+ </Tab>
+</Tabs>
+
+## How It Works
+
+This script demonstrates a practical workflow for extracting LinkedIn connection data:
+
+1. **Session Persistence** - Manually log into LinkedIn through the VM once, and the VM saves your session so the agent appears as your regular browsing.
+2. **Navigation** - The script navigates to your LinkedIn connections page using your saved authenticated session.
+3. **Data Extraction** - For each contact, the agent clicks their profile, extracts name/role/company/URL, and navigates back to repeat.
+4. **Python Processing** - Python parses the agent's responses, validates data, and writes to CSV incrementally to preserve progress.
+5. **Output Files** - Generates a CSV with contact data and a text file with direct messaging URLs.
+
+## Next Steps
+
+- Learn more about [Cua computers](/computer-sdk/computers) and [computer commands](/computer-sdk/commands)
+- Read about [Agent loops](/agent-sdk/agent-loops), [tools](/agent-sdk/custom-tools), and [supported model providers](/agent-sdk/supported-model-providers/)
+- Experiment with different [Models and Providers](/agent-sdk/supported-model-providers/)
+- Adapt this script for other platforms (Twitter/X, email extraction, etc.)