Remove letters from text

Remove Letters keeps everything in your text except alphabetic letters. It is a focused way to isolate order numbers, reference codes, punctuation, separators, emoji, currency marks, and layout without manually constructing a fragile character list. The operation recognizes Unicode letters rather than only English A through Z, so Latin, Greek, Cyrillic, Arabic, CJK, and other letter scripts are handled consistently. Characters that are not letters remain exactly where they appeared, and the response reports both the filtered result and the number of letters removed.

Remove letters without disturbing the rest of the text

Paste or send a string in the required text field. The capability scans it by Unicode code point and removes characters classified in a Unicode Letter category. Everything else is copied to the result in its original order.

That means decimal digits remain digits, punctuation such as commas, colons, dashes, quotes, and brackets stays in place, whitespace and line breaks are retained, and symbols such as currency signs, mathematical operators, emoji, and pictographs are preserved. For example, an input containing an order label, an identifier, and a time can lose the descriptive words while retaining the identifier separators and clock punctuation.

This is intentionally different from extracting numbers alone: a number extractor may normalize signs, decimals, or grouping, whereas this operation makes only one decision for each character—whether it is a letter. An empty string is valid and produces an empty result with a zero count, because it is real text with nothing to remove.

Unicode-aware behavior across languages and scripts

The definition of a letter is based on the Unicode Letter general category, not a hand-written English alphabet range. As a result, accented Latin letters, Greek and Cyrillic characters, Arabic letters, Hebrew letters, East Asian ideographs, and letters from many other writing systems are removed. Uppercase, lowercase, titlecase, modifier letters, and other Unicode letter subcategories are all included.

This avoids the common failure of an A–Z regular expression that leaves non-English words behind. At the same time, the algorithm does not normalize the input before filtering. A combining accent stored as its own code point is a Unicode mark rather than a letter, so it is preserved when its base letter is removed; a precomposed accented character such as é is itself a letter and is removed as one character.

That literal code-point policy best matches the promise to keep non-letters unchanged, and it prevents invisible normalization from rewriting punctuation or symbols. If your workflow needs diacritic cleanup, ASCII transliteration, or normalization, apply a dedicated capability before or after this operation. The filter never changes the order of retained characters and never collapses repeated spaces or separators.

Use the retained structure in codes and data pipelines

This tool is most useful when letters are noise but the structure surrounding numbers matters. A support team can strip prose from copied ticket references and retain digits, hashes, slashes, and dashes for downstream inspection. Finance and operations pipelines can expose amounts, percent signs, decimal punctuation, and currency symbols while removing nearby labels, making a later purpose-built parser easier to configure.

Because spaces and line breaks remain, the result may contain empty-looking regions where words were removed; that is deliberate evidence of the original layout, not accidental padding. If you want compact output, pass the result to a whitespace normalization capability afterward rather than combining unrelated behavior here. Likewise, this capability does not validate whether retained digits form a phone number, date, price, or account code.

It only removes letters and reports how many it removed. Keeping that contract narrow makes the output predictable and composable. Validate the semantic meaning of any retained code separately, especially where leading zeros, decimal separators, or signs are significant.