Extract URLs From Text Online

The Challenge of Extracting URLs from Raw Text

In modern digital workflows, links are embedded across disparate sources: support tickets, Slack/Discord chat histories, marketing documents, email dumps, and PDF reports. Extracting these URLs by hand is tedious and error-prone.

However, building an automated URL extractor is deceptively complex due to the informal way humans write in natural language:

• Trailing Punctuation: Sentences often end with URLs followed immediately by punctuation: "Check our site at https://example.com." or "Look at (https://example.com),". A naive regular expression will capture the trailing period or comma, creating an invalid 404 URL.
• Balanced Parentheses: Certain websites—notably Wikipedia—contain balanced parentheses inside valid URLs, such as https://en.wikipedia.org/wiki/Rust_(programming_language). Simply stripping all parentheses corrupts the link.
• Tracking Bloat: Shared URLs frequently contain massive query strings loaded with marketing trackers (?utm_source=twitter&utm_medium=social&utm_campaign=launch&fbclid=IwAR...). These trackers clutter databases and break deduplication algorithms.

Stringra's Robust Extraction Pipeline

Our client-side extractor uses a multi-stage tokenization algorithm:

1. Scheme & Authority Pattern Matching: Scans for valid URL URI schemes (https://, http://, ftp://) or www-prefixed domain boundaries.
2. Punctuation Boundary Cleaning: Examines the trailing edge of each matched token. Strips unmatched closing brackets, periods, commas, semicolons, and quotes.
3. URL Normalization & Parameter Stripping: If the "Strip query params" option is selected, the URL is parsed via the browser's native URL Web API to drop search params while preserving clean pathnames and anchors.
4. Domain Extraction: In "Domains only" mode, extracts only the normalized host (e.g. github.com), stripping protocol and sub-paths for domain-level backlink auditing.

Extracting URLs in Popular Programming Languages

Python (re + urllib.parse):

import re
from urllib.parse import urlparse

def extract_urls(text: str, strip_params: bool = False) -> list[str]:
    # Regex matching URL pattern
    pattern = r'https?://[^\s<>"]+|www\.[^\s<>"]+'
    matches = re.findall(pattern, text)
    
    cleaned = []
    for url in matches:
        # Strip trailing punctuation common in natural language
        url = re.sub(r'[\.,;:\!\?\)]+$', '', url)
        if strip_params:
            parsed = urlparse(url)
            url = f"{parsed.scheme}://{parsed.netloc}{parsed.path}"
        cleaned.append(url)
        
    return list(dict.fromkeys(cleaned)) # Deduplicate preserving order

JavaScript / TypeScript:

function extractUrls(text: string): string[] {
  const urlRegex = /\b(?:https?:\/\/|www\.)[^\s<>"'()]+(?:\([^\s<>"'()]+\)|[^\s<>"'()]*[^\s<>"'().,:;!?])/gi;
  const matches = text.match(urlRegex) || [];
  
  return [...new Set(matches.map(url => {
    return url.startsWith('www.') ? `https://${url}` : url;
  }))];
}

Unix Shell (grep + awk):

# Extract and sort unique URLs from a text file
grep -o -E 'https?://[^ ]+' access.log | sort -u > urls.txt

Privacy & Security: Zero Server Transmission

Chat logs, internal ticket exports, and developer documentation frequently contain sensitive staging URLs, private GitHub repository links, or pre-signed AWS S3 bucket tokens.

Stringra executes all parsing 100% locally in your browser. Your documents are never sent across a network, never logged, and never stored on any remote cloud server.