Invisible zero-width characters copied from rich-text editors secretly trigger UCS-2 encoding and multi-segment billing; sanitize input strings before dispatch.
One of the most insidious bugs in messaging engineering involves zero-width spaces (U+200B), zero-width joiners (U+200D), and byte-order marks (U+FEFF). These characters are completely invisible in text editors and browser forms, yet they trigger immediate UCS-2 encoding and segment explosion.
1 The Zero-Width Segment Bomb Explained
Consider a standard 100-character welcome message. If a copywriter inadvertently pastes a hidden Zero-Width Space (U+200B) from Notion or Google Docs:
- The text appears completely normal on screen.
- Carrier encoding engines detect Unicode and switch from GSM-7 to UCS-2.
- Maximum segment capacity drops from 160 to 70 characters.
- The 100-character message now costs 2 billable segments instead of 1.
2 Sanitization Regex Implementation
Deploy pre-send sanitization to strip zero-width artifacts:
Python: re.sub(r'[\u200B-\u200D\uFEFF]', '', text)
JavaScript: text.replace(/[\u200B-\u200D\uFEFF]/g, '')
Frequently Asked Questions • Character Encoding & Economics
Yes. Malicious actors can paste zero-width characters into user profile fields that are later concatenated into outgoing SMS notifications, intentionally inflating your carrier billing.