Text Splitter — Split by punctuation, word, phrase, or pattern

1. What does this tool do

This free online text splitter splits text by punctuation, word, phrase, or custom regex pattern. You get an ordered list of chunks that you can copy as newline- or comma-separated. Options let you trim each chunk and omit empty segments. Use it to break text into parts for data prep, parsing, or further processing. No sign-up, no upload; all processing runs in your browser.

2. How to use it

Quick start: Choose a mode (Punctuation, Word, Phrase, or Pattern), enter a delimiter or pattern if needed, paste your text, then click Split. Copy the chunks in your preferred format.

1. Select mode — Punctuation, Word, Phrase, or Pattern.
2. Punctuation mode — Choose All punctuation (any Unicode punctuation) or Selected punctuation and check the characters to split on (e.g. period, comma, semicolon).
3. Phrase mode — Enter the exact delimiter string (e.g. " and ", "---"). Choose whether to Remove the delimiter, Keep with next chunk, or Keep with previous chunk. Empty delimiter is not allowed.
4. Pattern mode — Enter one or more JavaScript regex patterns and optional flags (g, i, m). You can add multiple patterns; they are applied in order: the first pattern splits the text, the second splits each of those chunks, and so on. The result is one flat list of chunks, then trim/omit empty apply. Example: split by double newline (paragraphs), then by sentence-ending punctuation to get a list of sentences. The same validation and length limits as the Regex Cleaner apply.
5. Options — Check Trim each chunk and Omit empty chunks to clean the result (both default on).
6. Enter or paste text — Type or paste into the input area.
7. Click Split — The tool splits the text and shows the chunk count and list.
8. Copy or Export — Copy as Newline or Comma, or click Export CSV to download a CSV file (optionally with an index column).

3. How it works

• Punctuation — All: splits on one or more Unicode punctuation characters (\p{P}). Selected: splits only on the punctuation characters you check (e.g. . , !). Chunks are the segments between those runs.
• Word — Splits on whitespace (same logic as the Text Tokenizer’s Words mode). Leading/trailing punctuation on each word is not stripped for the split boundaries; trim/omit empty still apply.
• Phrase — Splits on the literal delimiter string you provide. No regex; the delimiter is used as-is. You can Remove the delimiter from output, Keep with next chunk (each chunk after the first starts with the delimiter), or Keep with previous chunk (each chunk except the last ends with the delimiter). Trim applies to the whole chunk.
• Pattern — Builds a regex from your pattern and flags, then uses String.prototype.split(regex). You can add multiple patterns; they are applied in sequence (first pattern splits the input, second splits each resulting chunk, etc.), producing one flat list of chunks before trim/omit empty. Pattern and input length limits match the Regex Cleaner to reduce ReDoS and abuse.

Export CSV — After splitting, you can download the chunks as a CSV file with one row per chunk. Optionally include a 1-based index column. Fields are escaped per RFC 4180 (commas, newlines, and double quotes).

All computation runs in your browser. No data is sent to any server.

4. Use cases & examples

• Data preparation — Split CSV-like or log lines by a delimiter (phrase or pattern).
• Sentence or clause extraction — Split by punctuation to get rough segments.
• Word lists — Use Word mode to get a list of words (with trim/omit empty).
• Custom parsing — Use Pattern mode with a regex when you need flexible split rules.

Example (Phrase mode, delimiter " and ", Remove):
Input: "apples and oranges and bananas"
Chunks: apples, oranges, bananas

Example (Phrase mode, Keep with next chunk):
Input: "a|b|c", delimiter |
Chunks: a, |b, |c

Example (Punctuation mode, All):
Input: "Hello. World! How are you?"
Chunks (with trim/omit empty): Hello, World, How are you

Example (Punctuation mode, Selected: period and comma):
Input: "One. Two, three."
Chunks: One, Two, three

5. Limitations & known constraints

• Input cap — Maximum 100,000 characters. Larger input returns an error.
• Pattern mode — Pattern length limited to 500 characters; same validation as Regex Cleaner. JavaScript regex syntax only.
• Client-side only — No server; processing runs in the browser. Very large inputs may cause brief UI lag on slower devices.
• Phrase mode — Delimiter is case-sensitive and literal; no regex.