# Convert Number Words to Digits Convert numbers written in English words into digits — inside prose, line by line, or as a list. Handles decimals, fractions, ordinals, lakh/crore and long scale. ## Run it - **CLI:** `gizza tool number-words-to-digits "We shipped twenty-five units and one hundred and two spares."` - **Web:** https://gizza.ai/tools/number-words-to-digits/ - **Agents:** machine-readable descriptor (parameters JSON Schema) at https://gizza.ai/tools/number-words-to-digits/tool.json ## Inputs - `input` — Text or number words _(field)_ - `mode` — What to return _(field)_ - `separator` — Thousands separator _(field)_ - `scale` — Billion / trillion reading _(field)_ - `ordinals` — Ordinal words _(field)_ - `fractions` — Read half and quarter as fractions _(field)_ - `digit_sequences` — Read digit runs (nine one one → 911) _(field)_ ## Output - Digits (text) ## Query parameters Open the tool pre-filled and auto-run via URL: - `input` — Text or number words - `mode` — What to return - `separator` — Thousands separator - `scale` — Billion / trillion reading - `ordinals` — Ordinal words - `fractions` — Read half and quarter as fractions - `digit_sequences` — Read digit runs (nine one one → 911) Example: `https://gizza.ai/tools/number-words-to-digits/?input=We%20shipped%20twenty-five%20units%20and%20one%20hundred%20and%20two%20spares.&mode=replace&separator=none&scale=short&ordinals=cardinal&fractions=true&digit_sequences=true` --- ## Turn spelled-out numbers into digits Paste anything that contains numbers written in English words — a sentence, a transcript, a column of phrases from a spreadsheet — and get the digits back. The default mode rewrites the text in place and leaves every other word alone, so a paragraph stays a paragraph. Everything runs in your browser; the text you paste is never uploaded. ### Worked example Using the **Numbers in prose** preset: - Text or number words: `We shipped twenty-five units and one hundred and two spares.` - What to return: `replace` - Thousands separator: `none` - Ordinal words: `cardinal` The result is: `We shipped 25 units and 102 spares.` Switch **What to return** to `value`, paste `one million two hundred fifty thousand`, and pick the comma separator to get `1,250,000` instead. ### What the controls mean **What to return** picks the shape of the output. `replace` (default) rewrites numbers where they appear and keeps the surrounding prose. `value` treats every non-empty line as one complete number phrase and reports an error naming the first word that does not belong — useful for cleaning a column of data. `extract` throws the prose away and returns just the numbers, one per line. **Thousands separator** only affects how digits are rendered: `none` keeps the output machine-readable, while comma, space and underscore group by threes. **Billion / trillion reading** switches between the short scale (a billion is 10⁹) and the long scale used in much of continental Europe (a billion is 10¹², a trillion 10¹⁸). `milliard` is always 10⁹ in both, and the Indian `lakh` (100,000) and `crore` (10,000,000) are always accepted. **Ordinal words** decides what `twenty-first` becomes: `21`, `21st`, or nothing at all. **Read half and quarter as fractions** turns `one and a half` into `1.5`; turn it off when those words are ordinary English in your text. **Read digit runs** is off by default because it changes meaning: with it on, `one two three` becomes `123` rather than `1 2 3`. ### What it understands Units, teens and tens (`seven`, `nineteen`, `forty`), hyphenated compounds (`twenty-five`), `hundred` and scale words up to `decillion`, the connector `and` only where English really uses it (`one hundred and forty-two`, but `one and two` stays two numbers), decimal words (`five point forty-seven` → `5.47`, `one point five million` → `1500000`), fraction words (`three quarters` → `0.75`, `half a million` → `500000`), and negatives (`minus one hundred`, `negative twenty one`). ### Limits and edge cases - Input is capped at **200,000 characters**; longer text is rejected with a clear message. - Arithmetic is exact 128-bit decimal, so values must fit in roughly **±1.7 × 10³⁸**. `one decillion` (10³³) works; a long-scale `one septillion` (10⁴²) reports "number is too large" rather than losing precision to floating point. - Only `half` and `quarter` are treated as fractions — `third`, `fifth` and `eighth` are also ordinal words, and thirds have no exact decimal form. - English only. Roman numerals, currency phrases such as `one dollar and fifty cents`, and multiplicative phrases such as `six sixes` are not parsed. - `value` mode is deliberately strict: a stray word on a line is an error, not a silent skip. Use `extract` or `replace` for messy text. ## FAQ
Why did "one and two" not become 12 or 1.2? `and` is only treated as part of a number where English actually uses it — after `hundred` or a scale word (`one hundred and forty-two`, `one thousand and one`), or before a fraction (`one and a half`). Everywhere else it stays an ordinary conjunction, so `one and two` converts to `1 and 2`.
How do I convert a whole column of phrases at once? Paste one phrase per line and set **What to return** to `value`. Each line is converted independently and the results come back in the same order, so you can paste the output straight back into a spreadsheet. Blank lines are skipped.
Does it understand lakh and crore? Yes. `three lakh` is `300000` and `two crore fifty lakh` is `25000000`, in both the short-scale and long-scale settings — those words never collide with `million`/`billion`. Indian *output* grouping (`12,34,567`) is not offered here; the separator setting groups by threes.
What happens to ordinals like "twenty-first"? With the default `cardinal` setting it becomes `21`. Choose `suffix` for `21st`, which is what you usually want when rewriting dates in prose. Choose `ignore` to stop ordinal words ending a number phrase — note that in replace mode `twenty-first` then shows as `20-first`, because only the cardinal part converts.
Is my text sent anywhere? No. The converter is compiled to WebAssembly and runs entirely in your browser — there is no server call, no logging and no network access of any kind. The same code is available offline through the command line.
## Related tools - [Extract Action Items from Meeting Notes](https://gizza.ai/tools/action-item-extractor/): Extract action items, owners, and decisions from meeting notes or daily notes with deterministic rules. Markdown checklist or JSON, private in-browser. - [Add Line Numbers](https://gizza.ai/tools/add-line-numbers/): Add line numbers to every line of text online, like nl or cat -n — custom start, step, separator, and alignment. Free and private, runs in your browser. - [ANSI Log Renderer](https://gizza.ai/tools/ansi-log-renderer/): Paste ANSI-colored terminal output or CI logs and render them as HTML, or strip escape codes to plain text. Handles 16-color, 256-color, and truecolor SGR codes. - [API response diff](https://gizza.ai/tools/api-response-diff/): Compare two JSON API responses and see only the meaningful changes: ignore request ids, timestamps and UUIDs, match arrays by key or as sets, export a JSON Patch. - [Rust AST Diff](https://gizza.ai/tools/ast-diff/): Compare two Rust source snippets structurally. The tool parses both files, canonicalizes their ASTs, and ignores formatting, whitespace, and comments.