I needed to download a bunch of Uber receipts for my business expenses, and while Uber provides a way to manually download receipt PDFs one at a time through their website, there’s no batch download option. I asked Claude to help me write a Python script that would automate this process.
The script uses Playwright to control an existing browser session (to avoid having to handle login credentials), navigates through Uber’s trip history, and downloads each receipt as a PDF. The trickiest parts were handling the modal dialogs that appear when requesting receipts, and extracting the correct cost from the trip details page for the filename.
After some back-and-forth testing with Claude using its access to a test browser, we got it working reliably. The script now downloads receipts with filenames in a consistent format (YYYY-MM-DD-26.44USD-tripid.pdf), making them easy to organize and process for accounting.
I used the claude.ai desktop app with various MCP-servers and its artifact feature to work on this. Artifacts now use line-by-line editing to make changes, so the Claude Web interface and desktop client doesn’t necessarily re-output your code with every edit. This is about as big as I can get a Claude Sonnet 3.7-written program to get before it starts running out of context or getting confused.
Here’s the latest, maintained code for the script
To use it, you need the Python jack-of-all-trades, uv, installed.
Run your browser with remote debugging enabled (e.g., chromium
--remote-debugging-port=9222), then run the script with your desired
options:
# Download all available receipts
./uber-receipt-downloader.py --all
# Download receipts from a specific date range
./uber-receipt-downloader.py --start-date 2024-01-01 --end-date 2024-03-31
# Download receipts from the last 30 days
./uber-receipt-downloader.py --days 30
As the script mentions, Stack Overflow has instructions for getting Chrome running with remote debugging working on MacOS and Windows.
#!/usr/bin/env -S uv run
# /// script
# requires-python = ">=3.12"
# dependencies = [
# "playwright",
# ]
# ///
"""
Uber Receipt Downloader
This script connects to an existing Chrome-compatible browser instance and downloads
Uber receipts as PDFs from the Uber riders website.
Launch chrome with: chrome --remote-debugging-port=9222
Some instructions on how to do that on [MacOS and Windows](https://stackoverflow.com/questions/51563287/how-to-make-chrome-always-launch-with-remote-debugging-port-flag).
"""
import asyncio
import os
import re
import time
import argparse
from datetime import datetime, timedelta
from typing import List, Optional, Dict, Any
from playwright.async_api import async_playwright
# Configuration
CDP_URL = "http://localhost:9222" # CDP URL for Brave browser
DOWNLOAD_DIR = os.path.expanduser("~/Downloads/uber_receipts") # Directory to save receipts
# Set to None to fetch all trips automatically
TRIP_IDS = None
class UberReceiptDownloader:
def __init__(self, cdp_url: str, download_dir: str):
self.cdp_url = cdp_url
self.download_dir = download_dir
self.browser = None
self.context = None
self.page = None
self.playwright = None
async def connect_to_browser(self):
"""Connect to existing Brave browser instance over CDP"""
try:
self.playwright = await async_playwright().start()
print(f"Connecting to Brave browser at {self.cdp_url}")
self.browser = await self.playwright.chromium.connect_over_cdp(self.cdp_url)
# Get the default browser context
self.context = self.browser.contexts[0] if self.browser.contexts else await self.browser.new_context(
ignore_https_errors=True, # Ignore SSL certificate errors
accept_downloads=True, # Enable file downloads
)
print("Successfully connected to Brave browser")
except Exception as e:
print(f"Error connecting to Brave browser at {self.cdp_url}: {e}")
print("Make sure Brave is running with the --remote-debugging-port=9222 flag")
raise
async def extract_cost(self, page):
"""Extract the cost value from the trip page"""
try:
# Approach 1: Find all trip detail divs and look for the one with a $ symbol
cost_divs = await page.query_selector_all('div[data-baseweb="block"][class*="css-iMyxrY"]')
for div in cost_divs:
text = await div.inner_text()
if '$' in text:
# Use regex to extract just the number part
import re
price_match = re.search(r'\$(\d+\.\d+)', text)
if price_match and price_match.group(1):
return price_match.group(1)
# Approach 2: Try to find the Tag icon and get its parent div text
tag_icon = await page.query_selector('svg[title="Tag"]')
if tag_icon:
# Get the parent element that contains the price
parent_element = await page.query_selector('svg[title="Tag"] + div')
if parent_element:
price_element = await parent_element.query_selector('p')
if price_element:
price_text = await price_element.inner_text()
# Extract the number part
import re
price_match = re.search(r'\$(\d+\.\d+)', price_text)
if price_match and price_match.group(1):
return price_match.group(1)
# If we get here, we couldn't find the cost
print("Warning: Could not extract cost from page")
return "unknown"
except Exception as e:
print(f"Error extracting cost: {e}")
return "unknown"
async def fetch_trip_ids(self, start_date: Optional[datetime] = None, end_date: Optional[datetime] = None) -> List[str]:
"""Fetch trip IDs from the Uber trips page within a date range"""
if not self.browser or not self.context:
raise ValueError("Browser not connected. Call connect_to_browser() first.")
if not self.page:
self.page = await self.context.new_page()
try:
# Navigate to trips page
trips_url = "https://riders.uber.com/trips"
print(f"Navigating to {trips_url}")
await self.page.goto(trips_url, wait_until='networkidle', timeout=30000)
# Wait for the page to load with trip elements
print("Waiting for trip elements to load...")
await self.page.wait_for_selector('div[href^="https://riders.uber.com/trips/"]', state='visible', timeout=10000)
# Extract trip IDs and dates using JavaScript evaluation
trips_with_dates = await self.page.evaluate('''
() => {
// Look for divs with href attributes containing trip IDs
const tripElements = document.querySelectorAll('div[href^="https://riders.uber.com/trips/"]');
const trips = [];
for (const element of tripElements) {
const href = element.getAttribute('href');
let tripId = null;
if (href) {
const match = href.match(/\\/trips\\/([^\\/?]+)/);
if (match && match[1]) {
tripId = match[1];
}
}
// Try to find the date text
let dateText = null;
const dateElement = element.querySelector('div[data-baseweb="block"] div');
if (dateElement) {
dateText = dateElement.innerText;
}
if (tripId) {
trips.push({ id: tripId, dateText: dateText });
}
}
return trips;
}
''')
print(f"Found {len(trips_with_dates)} trip entries on the first page")
# Check if we need to load more trips (click "More" button if it exists)
try:
more_button_selector = 'button:has-text("More")'
while await self.page.is_visible(more_button_selector, timeout=5000):
print("Found 'More' button, clicking to load additional trips...")
await self.page.click(more_button_selector)
# Wait for loading to complete
await asyncio.sleep(2)
# Extract additional trip IDs
new_trips_with_dates = await self.page.evaluate('''
() => {
// Same extraction logic as before
const tripElements = document.querySelectorAll('div[href^="https://riders.uber.com/trips/"]');
const trips = [];
for (const element of tripElements) {
const href = element.getAttribute('href');
let tripId = null;
if (href) {
const match = href.match(/\\/trips\\/([^\\/?]+)/);
if (match && match[1]) {
tripId = match[1];
}
}
// Try to find the date text
let dateText = null;
const dateElement = element.querySelector('div[data-baseweb="block"] div');
if (dateElement) {
dateText = dateElement.innerText;
}
if (tripId) {
trips.push({ id: tripId, dateText: dateText });
}
}
return trips;
}
''')
if len(new_trips_with_dates) > len(trips_with_dates):
print(f"Loaded more trips, now found {len(new_trips_with_dates)} total")
trips_with_dates = new_trips_with_dates
else:
print("No new trips found after clicking 'More', stopping")
break
except Exception as e:
print(f"Note: Could not find or click 'More' button: {e}")
# Filter by date if requested
trip_ids = []
if start_date or end_date:
print(f"Filtering trips between {start_date.strftime('%Y-%m-%d') if start_date else 'earliest'} and {end_date.strftime('%Y-%m-%d') if end_date else 'latest'}")
# Process each trip to check date
for trip in trips_with_dates:
trip_id = trip['id']
date_text = trip['dateText']
# Skip if no date text available
if not date_text:
trip_ids.append(trip_id)
continue
# Try to extract date from date_text
try:
# Common patterns in Uber's interface:
# "Mar 6 • 2:25 PM"
# "March 6 • 2:25 PM"
import re
# Look for month names
months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun",
"Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
"January", "February", "March", "April", "May", "June",
"July", "August", "September", "October", "November", "December"]
month_pattern = "|".join(months)
match = re.search(rf'({month_pattern}) (\d+)', date_text)
if match:
month, day = match.groups()
# Use current year as Uber typically only shows month and day
# This is a limitation - for trips more than a year old
year = datetime.now().year
# Try to parse with both full and abbreviated month names
trip_date = None
for fmt in ["%B %d %Y", "%b %d %Y"]:
try:
trip_date = datetime.strptime(f"{month} {day} {year}", fmt)
break
except ValueError:
continue
# Check if date is in range
if trip_date:
if start_date and trip_date < start_date:
continue
if end_date and trip_date > end_date:
continue
# Include the trip if we couldn't parse the date or it's in range
trip_ids.append(trip_id)
except Exception as e:
# If date parsing fails, include the trip to be safe
print(f"Error parsing date for trip {trip_id}: {e}")
trip_ids.append(trip_id)
else:
# If no date filtering, include all trip IDs
trip_ids = [trip['id'] for trip in trips_with_dates]
# Remove duplicates while preserving order
seen = set()
unique_trip_ids = [x for x in trip_ids if not (x in seen or seen.add(x))]
print(f"Selected {len(unique_trip_ids)} trips after date filtering")
return unique_trip_ids
except Exception as e:
print(f"Error fetching trip IDs: {e}")
return []
async def download_receipt(self, trip_id: str) -> Optional[str]:
"""Download a receipt for a specific trip ID"""
if not self.browser or not self.context:
raise ValueError("Browser not connected. Call connect_to_browser() first.")
try:
if not self.page:
self.page = await self.context.new_page()
# Navigate to the trip page
trip_url = f"https://riders.uber.com/trips/{trip_id}"
print(f"Navigating to {trip_url}")
await self.page.goto(trip_url, wait_until='networkidle', timeout=30000)
# Click "View Receipt" button - use known selector from our testing
view_receipt_selectors = [
'button[data-tracking-name="view-receipt-link"]', # This worked in our testing
':text("View Receipt")',
'button:has-text("View Receipt")',
'[data-test="view-receipt-button"]'
]
receipt_clicked = False
for selector in view_receipt_selectors:
try:
# Check if the selector exists and is visible
visible = await self.page.is_visible(selector, timeout=2000)
if visible:
print(f"Found 'View Receipt' button with selector: {selector}, clicking...")
await self.page.click(selector)
receipt_clicked = True
break
except Exception:
continue
if not receipt_clicked:
print("Could not find 'View Receipt' button with any of the tried selectors")
# One last attempt - try to find it by looking at all buttons
try:
all_buttons = await self.page.query_selector_all('button')
for button in all_buttons:
text = await button.inner_text()
if "receipt" in text.lower() or "view" in text.lower():
print(f"Found button with text: {text}, clicking...")
await button.click()
receipt_clicked = True
break
except Exception as e:
print(f"Error in last attempt to find view receipt button: {e}")
if not receipt_clicked:
raise Exception("Could not find or click 'View Receipt' button")
# Wait for the popup dialog to appear
print("Waiting for receipt popup dialog...")
await asyncio.sleep(1) # Brief pause to ensure modal is fully loaded
# Click "Download PDF" button within the popup
download_pdf_selectors = [
':text("Download PDF")', # This worked in our testing
'text="Download PDF"',
'button:has-text("Download PDF")',
'[data-test="download-pdf-button"]'
]
pdf_button_found = False
for selector in download_pdf_selectors:
try:
# Check if the selector exists and is visible
visible = await self.page.is_visible(selector, timeout=2000)
if visible:
print(f"Found 'Download PDF' button with selector: {selector}")
pdf_button_found = True
# Set up download event handler
download_path = None
async with self.page.expect_download() as download_info:
# Now click the button inside the expect_download context manager
await self.page.click(selector)
print("Waiting for download to start...")
# Wait for the download to start with a timeout
try:
download = await asyncio.wait_for(download_info.value, 10.0)
# Extract the cost from the page before clicking Download
cost = await self.extract_cost(self.page)
# Get date from the page if possible
try:
date_element = await self.page.query_selector('div[data-baseweb="block"] div[data-baseweb="typo-labellarge"]')
date_text = await date_element.inner_text() if date_element else None
if date_text:
# Extract date from various formats that might appear
import re
# Try to find patterns like:
# "2:28 PM, Thursday March 6 2025"
# "Mar 6 • 2:25 PM"
# Look for month names
months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun",
"Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
"January", "February", "March", "April", "May", "June",
"July", "August", "September", "October", "November", "December"]
month_pattern = "|".join(months)
date_patterns = [
# Pattern like "March 6 2025"
rf'({month_pattern}) (\d+) (\d4)',
# Pattern like "Mar 6 • 2:25 PM" (year will be current year)
rf'({month_pattern}) (\d+)',
]
date_formatted = None
for pattern in date_patterns:
date_match = re.search(pattern, date_text)
if date_match:
try:
if len(date_match.groups()) == 3: # Full date with year
month, day, year = date_match.groups()
date_str = f"{month} {day} {year}"
# Try to parse with both full and abbreviated month names
for fmt in ["%B %d %Y", "%b %d %Y"]:
try:
date_obj = datetime.strptime(date_str, fmt)
date_formatted = date_obj.strftime("%Y-%m-%d")
break
except ValueError:
continue
else: # Date without year
month, day = date_match.groups()
# Use current year
year = datetime.now().year
date_str = f"{month} {day} {year}"
# Try to parse with both full and abbreviated month names
for fmt in ["%B %d %Y", "%b %d %Y"]:
try:
date_obj = datetime.strptime(date_str, fmt)
date_formatted = date_obj.strftime("%Y-%m-%d")
break
except ValueError:
continue
except Exception as e:
print(f"Error parsing date: {e}")
if date_formatted:
break
if not date_formatted:
# If no date could be parsed, use current date
date_formatted = datetime.now().strftime("%Y-%m-%d")
else:
date_formatted = datetime.now().strftime("%Y-%m-%d")
except Exception:
# If date extraction fails, use current date
date_formatted = datetime.now().strftime("%Y-%m-%d")
# Generate a filename with date, cost and trip ID
filename = f"{date_formatted}-{cost}USD-{trip_id}.pdf"
download_path = os.path.join(self.download_dir, filename)
# Create directory if it doesn't exist
os.makedirs(self.download_dir, exist_ok=True)
# Save the file
await download.save_as(download_path)
print(f"Downloaded receipt to: {download_path}")
except asyncio.TimeoutError:
print("Download didn't start within the timeout period")
# Wait for a moment to ensure download completes or dialog closes
await asyncio.sleep(2)
# Close the popup by clicking outside or on the X button
try:
# Try specific close button first
close_button_selectors = [
'button[aria-label="Close"]',
'.ReactModalPortal button',
':text("×")'
]
for close_selector in close_button_selectors:
if await self.page.is_visible(close_selector, timeout=1000):
await self.page.click(close_selector)
print("Closed receipt popup dialog")
break
else:
# If no close button found, try pressing Escape
await self.page.keyboard.press("Escape")
print("Closed receipt popup dialog with Escape key")
except Exception as e:
print(f"Error closing popup: {e} - continuing anyway")
return download_path
except Exception:
continue
if not pdf_button_found:
# Try to find any download-related element with text
try:
# Extract all text from the page
all_text = await self.page.evaluate('''
() => {
return document.body.innerText;
}
''')
if "download pdf" in all_text.lower():
# Try a more generic selector
print("Found 'Download PDF' text, trying generic selector")
selector = ':text("Download PDF")'
pdf_button_found = True
elif "download" in all_text.lower() and "pdf" in all_text.lower():
# Try clicking anywhere with download or PDF text
print("Found 'Download' and 'PDF' text separately, trying broader selector")
selector = ':text-matches("(?i)download|pdf")'
pdf_button_found = True
except Exception as e:
print(f"Error searching for download text: {e}")
if not pdf_button_found:
raise Exception("Could not find 'Download PDF' button")
except Exception as e:
print(f"Error downloading receipt for trip {trip_id}: {e}")
return None
async def download_multiple_receipts(self, trip_ids: Optional[List[str]] = None, start_date: Optional[datetime] = None, end_date: Optional[datetime] = None):
"""Download receipts for multiple trip IDs"""
# Connect to browser if not already connected
if not self.browser:
await self.connect_to_browser()
# If no trip IDs provided, fetch them
if not trip_ids:
print("No trip IDs provided, fetching from Uber...")
trip_ids = await self.fetch_trip_ids(start_date, end_date)
if not trip_ids:
print("No trips found.")
return []
print(f"Starting download of {len(trip_ids)} receipts...")
results = []
for i, trip_id in enumerate(trip_ids):
print(f"Downloading receipt {i+1}/{len(trip_ids)} - Trip ID: {trip_id}")
result = await self.download_receipt(trip_id)
results.append((trip_id, result))
# Small delay between downloads to avoid rate limiting
await asyncio.sleep(2)
return results
async def close(self):
"""Close browser connection and clean up resources"""
try:
if self.page:
await self.page.close()
if self.browser:
await self.browser.close()
if self.playwright:
await self.playwright.stop()
except Exception as e:
print(f"Error during cleanup: {e}")
async def main():
# Parse command line arguments
parser = argparse.ArgumentParser(description='Download Uber receipts as PDFs')
parser.add_argument('--trip-id', action='append', help='Specific trip ID(s) to download (can be used multiple times)')
parser.add_argument('--days', type=int, default=90, help='Number of days back to fetch trips (default: 90)')
parser.add_argument('--start-date', type=str, help='Start date for trip range (format: YYYY-MM-DD)')
parser.add_argument('--end-date', type=str, help='End date for trip range (format: YYYY-MM-DD, defaults to today)')
parser.add_argument('--output-dir', type=str, default=DOWNLOAD_DIR, help=f'Directory to save receipts (default: {DOWNLOAD_DIR})')
parser.add_argument('--cdp-url', type=str, default=CDP_URL, help=f'Chrome DevTools Protocol URL (default: {CDP_URL})')
parser.add_argument('--all', action='store_true', help='Download all available trips')
parser.add_argument('--test', action='store_true', help='Test cost extraction on a specific trip ID')
parser.add_argument('--test-trip-id', type=str, default="1003c9ae-bd1c-48eb-b751-e260c336f7fa", help='Trip ID to use for testing (default: a specific trip ID)')
args = parser.parse_args()
# Use arguments or defaults
cdp_url = args.cdp_url
download_dir = args.output_dir
# Initialize downloader
downloader = UberReceiptDownloader(cdp_url, download_dir)
try:
await downloader.connect_to_browser()
# Special test mode
if args.test:
test_trip_id = args.test_trip_id
print(f"Testing cost extraction on trip ID: {test_trip_id}")
# Navigate to the trip page
if not downloader.page:
downloader.page = await downloader.context.new_page()
trip_url = f"https://riders.uber.com/trips/{test_trip_id}"
print(f"Navigating to {trip_url}")
await downloader.page.goto(trip_url, wait_until='networkidle', timeout=30000)
# Extract the cost
cost = await downloader.extract_cost(downloader.page)
print(f"Extracted cost: {cost}")
return
# Process date arguments for normal operation
start_date = None
end_date = datetime.now()
if args.start_date:
try:
start_date = datetime.strptime(args.start_date, "%Y-%m-%d")
except ValueError:
print(f"Error: Invalid start date format. Please use YYYY-MM-DD. Using default.")
start_date = datetime.now() - timedelta(days=args.days)
else:
start_date = datetime.now() - timedelta(days=args.days)
if args.end_date:
try:
end_date = datetime.strptime(args.end_date, "%Y-%m-%d")
except ValueError:
print(f"Error: Invalid end date format. Please use YYYY-MM-DD. Using today.")
# Determine which trip IDs to use
trip_ids = args.trip_id if args.trip_id else TRIP_IDS
if trip_ids:
print(f"Using {len(trip_ids)} provided trip ID(s)")
results = await downloader.download_multiple_receipts(trip_ids)
elif args.all or TRIP_IDS is None:
# Fetch trip IDs automatically
print(f"Fetching trip IDs from {start_date.strftime('%Y-%m-%d')} to {end_date.strftime('%Y-%m-%d')}")
results = await downloader.download_multiple_receipts(
trip_ids=None,
start_date=start_date,
end_date=end_date
)
else:
print("No trip IDs provided and --all not specified. Please provide trip IDs or use --all.")
return
# Print results
successful = [trip_id for trip_id, path in results if path]
failed = [trip_id for trip_id, path in results if not path]
print("\nDownload Results:")
print(f"Successfully downloaded: {len(successful)}/{len(results)}")
if failed:
print(f"Failed downloads: {len(failed)}")
for trip_id in failed:
print(f" - {trip_id}")
if successful:
print(f"\nReceipts saved to: {download_dir}")
except Exception as e:
print(f"Error: {e}")
finally:
await downloader.close()
if __name__ == "__main__":
asyncio.run(main())