Advanced Data Sanitizer

The Technical Need for Data Sanitization

Logs, user feedback, and database dumps often contain Personally Identifiable Information (PII). Sharing or storing these raw records without cleaning them can violate privacy standards like GDPR, HIPAA, and CCPA. This utility helps you sanitize data locally by removing email addresses, phone numbers, and credentials before logging or exporting the data.

Data sanitization is not just a compliance requirement it is a fundamental security practice. In 2023 alone, data breaches exposed over 300 million records in the United States, many of which could have been mitigated by proper data handling practices. When developers copy log output into bug reports, share database exports with colleagues, or paste API responses into documentation, they risk inadvertently exposing sensitive information. A single email address, phone number, or API key in a shared log file can become the entry point for a phishing attack, social engineering campaign, or unauthorized access.

This tool addresses this problem by providing a comprehensive set of sanitization filters that run entirely in your browser. No data is uploaded to any server, which means the sensitive information in your logs, exports, and feedback forms never leaves your device. You can use it to prepare data for sharing, reduce the attack surface of your logs, and ensure that your data handling practices align with privacy regulations.

How the Sanitizer Works Technically

The sanitizer processes text input using a series of regular expression pattern matchers and string replacement functions. Each filter is designed to detect a specific type of sensitive data based on its characteristic format. Here is how each major filter works:

• Email Address Detection: The filter uses a regular expression that matches the standard email format: local-part@domain.tld. It identifies email addresses in the input text and replaces them with a redaction placeholder like [EMAIL_REDACTED]. The pattern handles common email formats including those with dots, plus signs, and subdomains in the local part.
• Phone Number Detection: The filter matches international and domestic phone number formats, including those with country codes, area codes, extensions, and various separator characters (dashes, dots, spaces, parentheses). It handles formats like +1-555-123-4567, (555) 123-4567, and 555.123.4567 ext. 100.
• IP Address Detection: The filter identifies both IPv4 addresses (four dot-separated decimal numbers) and IPv6 addresses (colon-separated hexadecimal groups). It replaces detected addresses with [IP_REDACTED] to prevent the exposure of network infrastructure details.
• Credit Card Number Detection: The filter matches the characteristic patterns of major credit card networks (Visa, Mastercard, American Express, Discover) based on their prefix ranges and digit lengths. It also validates detected numbers using the Luhn algorithm to reduce false positives, ensuring that only valid card number patterns are redacted.
• API Key and Token Detection: This filter identifies common patterns for API keys, bearer tokens, SSH keys, and other authentication credentials. It looks for characteristic prefixes like Bearer, sk_live_, pk_, AKIA (AWS access keys), and patterns that match base64-encoded tokens. Detected credentials are replaced with [TOKEN_REDACTED].
• Password and Credential Detection: The filter identifies inline password assignments like password=secretvalue, passwd=secretvalue, and similar patterns used in configuration files, connection strings, and log output. The entire key-value pair is replaced with [CREDENTIAL_REDACTED].
• Invisible Unicode Character Scrubber: This filter removes zero-width characters, non-joiners, soft hyphens, and other invisible Unicode characters that can be used to inject hidden content or obfuscate text. These characters are invisible to the human eye but can be interpreted by browsers and other software, potentially enabling content injection attacks.

All filters operate on the text string using pattern matching and replacement. The processing is deterministic and fast, even for large text inputs. The tool processes thousands of lines per second, making it practical for sanitizing large log files and database exports.

Sanitization Capabilities

The tool provides several categories of data cleaning capabilities:

• PII Redaction: Uses patterns to identify and mask email addresses, IP addresses, credit card numbers, SSNs, and phone numbers. The redaction is applied consistently across the entire document, ensuring that no instances are missed.
• Credential Scrubbing: Detects and removes API keys, bearer tokens, SSH keys, and inline password assignments. This is critical when sharing logs, configuration files, or API documentation that may accidentally contain embedded credentials.
• Invisible Character Removal: Strips zero-width characters, non-joiners, and other invisible Unicode characters that can be used for content injection or text obfuscation. This filter is particularly useful when processing user-submitted text that may contain hidden malicious content.
• Formatting Cleanup: Fixes extra spacing, trailing commas, and carriage return variations across systems. This is useful when combining text from multiple sources that use different line ending conventions.
• Encoding Repair: Fixes encoding errors (mojibake) caused by incorrect conversions between ISO-8859-1 and UTF-8. This occurs when text is copied between systems that use different default encodings, resulting in garbled characters like instead of an apostrophe.

Practical Use Cases

Data sanitization is relevant in many professional contexts:

• Bug Reports and Issue Tracking: When filing bug reports, developers often include log output or API responses as evidence. Running this data through the sanitizer ensures that no PII or credentials are exposed in the issue tracker, which may be accessible to a wide audience.
• Log Sharing and Collaboration: When troubleshooting production issues, you may need to share logs with colleagues, vendors, or support teams. Sanitizing the logs first removes sensitive information while preserving the technical details needed for diagnosis.
• Data Migration and Export: When exporting data from one system to another, you may need to strip PII to comply with data protection regulations. This tool lets you sanitize data locally before uploading it to a new system or sharing it with a third party.
• Training Data Preparation: When preparing datasets for machine learning or analytics, you often need to remove personal identifiers to comply with privacy regulations. The sanitizer can process text-based datasets to strip emails, phone numbers, and other PII.
• Customer Feedback Processing: User feedback often contains personal information that users inadvertently include in their messages. Sanitizing feedback before storing or analyzing it protects user privacy.
• Compliance Auditing: When auditing data handling practices, you can use this tool to demonstrate that your sanitization processes are working correctly and that sensitive data is being properly redacted.
• Security Incident Response: During security investigations, you may need to share log excerpts with external parties. Sanitizing the data first ensures that credentials and personal information are not exposed during the investigation.

Comparison with Alternative Approaches

There are several approaches to data sanitization, each with different tradeoffs:

• Server-Side Sanitization Services: Cloud-based services that process your data on their servers. While they may offer more sophisticated detection algorithms, they require you to upload sensitive data to a third party, which can be a compliance violation.
• Command-Line Tools: Utilities like sed, awk, and grep can be used to redact patterns from text files. These are powerful but require expertise with regular expressions and command-line syntax. They also run locally, which is good for privacy, but they lack the visual feedback and multiple filter options of a dedicated tool.
• Programming Libraries: Libraries like Python's faker or scrubadub provide programmatic sanitization. These are flexible but require coding knowledge and setup time. They are best suited for automated pipelines rather than ad-hoc sanitization tasks.
• Browser-Based Local Tools: This tool falls into this category. It provides a visual interface with multiple filter options, custom rule support, and immediate feedback, all running locally in your browser. It is the most accessible option for developers who need to sanitize data quickly without setup or data transmission.

Custom Replacement Rules

Beyond the built-in filters, the tool supports custom replacement rules. This feature lets you define your own pattern-replacement pairs for data patterns that are specific to your organization or use case. For example, you might want to redact:

• Internal employee ID numbers that follow a specific format
• Project codenames or internal product names
• Server hostnames and internal DNS names
• Custom identifiers used in your organization's systems

Each custom rule consists of a search pattern (which can be a regular expression) and a replacement string. The rules are applied in sequence after the built-in filters, giving you full control over what gets sanitized and how it is replaced.

Redaction Treatment Modes

The tool offers three different treatment modes for detected sensitive data:

• Redact Label: Replaces detected data with a descriptive placeholder like [EMAIL_REDACTED] or [PHONE_REDACTED]. This is the most common mode because it clearly indicates what type of data was removed while preserving the structure of the surrounding text.
• Character Masking: Replaces the sensitive data with asterisks or other mask characters while preserving the length and general structure. For example, an email address might become j***n@example.com. This is useful when you need to verify that data was present without revealing its actual content.
• SHA-256 Hash: Replaces the sensitive data with a cryptographic hash. This is useful when you need to correlate records across datasets without exposing the original values. The same input will always produce the same hash, allowing you to match records without revealing the underlying data.

Tips and Best Practices for Data Sanitization

• Review the Output: After sanitizing data, always review the output to verify that the correct data was redacted. No pattern-based detection is perfect, and some types of sensitive data may not match the built-in patterns.
• Use Custom Rules for Organization-Specific Data: If your data contains organization-specific identifiers that are not covered by the built-in filters, create custom rules to handle them. This ensures comprehensive coverage.
• Combine Multiple Filters: Enable all relevant filters for your use case. If you are sanitizing logs for sharing, enable email, phone, IP, credit card, and credential filters simultaneously. The tool applies all enabled filters in a single pass.
• Process Data Locally: Always run the sanitizer on your local machine. Never upload sensitive data to online sanitization services, as this defeats the purpose of protecting the data.
• Test with Sample Data: Before processing large datasets, test your sanitization configuration with a small sample to verify that the filters are working as expected.
• Keep the Original: Always retain the original, unsanitized data in a secure location. The sanitized version is for sharing; the original may be needed for auditing or investigation purposes.