Text Splitter
Paste a block of text, pick a split mode (chars, words, sentences, or delimiter), then break it into clean chunks. Copy each chunk on its own or grab all of them in order.
Split mode
Private by default
Everything runs in your browser. Your text is not sent to a server and is not logged.
How to use the text splitter
1. Paste or type your input
Drop the block you want to break apart into the input box. A long paragraph, a draft thread, an essay, a system log, or any block of prose all work. There is no minimum length, and there is no hard upper limit either: a few megabytes is fine.
2. Pick a split mode
Four modes cover the common jobs. By character count cuts every N characters, snapping to the nearest space before the limit so words stay whole. This is the default at 280 characters because that matches the X (Twitter) per-post limit. By word count splits every N words and keeps every word intact. By sentence splits at the next sentence boundary (. ! or ?). By delimiter splits whenever a string you choose appears in the input, which is handy for CSV-style blocks or log files.
3. Set the chunk size or delimiter
The relevant control appears below the mode selector. For character mode, type a chars-per-chunk value (default 280). For word mode, type a words-per-chunk value (default 100). For delimiter mode, type the exact string to split on (default ---). Sentence mode has no extra input.
4. Pick options, then split
Turn on Number each chunk if you want the splitter to prefix each output with 1/, 2/, 3/ (useful for thread numbering on social posts). Turn off Keep last chunk if shorter if a tiny trailing fragment is annoying you (the splitter will drop any final chunk below 25 percent of the target). Click Split. The output appears below as stacked chunk cards. Use Copy on any single card, or Copy all chunks to grab everything in order with one click.
When you need to split text
Splitting a block of text into smaller pieces is one of those jobs that crops up across very different workflows. Below are the cases where the splitter pays for itself, with the mode and chunk size that usually fit.
Twitter thread prep
X (Twitter) caps each post at 280 characters. Writers drafting a thread in a single document need to break that draft into per-post chunks before they paste back into the composer. Pick By character count, leave the default 280, and turn on Number each chunk. Each chunk card is exactly one ready-to-post tweet, in order.
SMS batching
Carrier-side SMS protocols typically split anything past 160 characters into multiple messages (or 70 characters for messages with emoji). When you are pasting a long template into an SMS-broadcast tool that does not auto-split, drop the text in here, set chars-per-chunk to 160, and copy the chunks across one by one to keep them in order.
AI prompt token-budgeting
When a prompt is too long for a single LLM call (context window full, or the model is dropping the tail), splitting the prompt into ordered chunks lets you feed them sequentially. Character mode at 4000 characters is a rough proxy for around 1000 tokens, which keeps each chunk well inside most context windows. Pair with sentence mode if you want each chunk to end on a clean boundary.
Log chunking
Log files come delimited by timestamps, line breaks, or fixed markers. Drop the log block in, set the delimiter to your timestamp prefix (or use --- if the file is already separator-delimited), and you get one chunk per log entry. From there, paste any single entry into a bug ticket without manual cleanup.
ESL reading and translation prep
Language learners and translators often prefer to work one sentence at a time so the brain can hold both the source and the target in working memory. Sentence mode produces one chunk per sentence, ready to paste into a translation table or flashcard tool.
Chunking long-form content for review
Editors reviewing a long draft sometimes want to send it back to the writer one section at a time, with comments per chunk. Delimiter mode (set to the writer's section marker, often ## or ---) splits the draft along the writer's own structure, no manual scrolling required.
Tweet-thread prep tip: 280 chars plus "1/" numbering
The single most common use of a text splitter is preparing a Twitter thread, and the right settings are not obvious until you have done it once or twice. Here is the pattern that produces a clean, ready-to-paste thread every time.
Settings: Split mode By character count, Chars per chunk 280, Number each chunk toggled on, Keep last chunk if shorter toggled on.
Why these numbers: 280 is the per-post character limit on X. Numbered output produces 1/, 2/, 3/ prefixes so readers see the thread order at a glance and so you can paste posts out of order if the composer asks you to retry. Keeping the short trailing chunk matters because it is usually your call-to-action ("follow for more" or a link), which you do not want the splitter to drop on the floor.
One catch: the 280-char limit counts a URL as 23 characters regardless of its real length, but the splitter counts every actual character. If you are including a long URL in a single post, leave at least 25 characters of headroom in that chunk.
Once you have the chunks, double check the count of each post matches what you expect with a quick paste into the Twitter character counter. It will flag any chunk over 280 (which can happen when a long word lands right at the boundary) before you start posting, so you can shave a word or re-split without losing the order.
Related tools
Frequently asked questions
What is the difference between the four split modes?
By character count cuts the text every N characters, snapping to the nearest whitespace so words are not chopped (use this for Twitter threads or SMS). By word count keeps every word whole and splits every N words. By sentence splits at . ! ? boundaries, one sentence per chunk. By delimiter splits whenever your custom string appears in the text (handy for CSV-like inputs or log files).
What is the default chunk size for Twitter threads?
280 characters, which is the X (Twitter) per-post limit. Set the split mode to By character count and leave the default. Turn on Number each chunk to prepend 1/, 2/, 3/ so readers can follow the order.
Does the text leave my browser?
No. The splitter runs entirely in your browser using JavaScript. Your input and the resulting chunks are never sent to our servers and are not logged. You can verify by opening the page, disconnecting from the internet, and seeing that splitting still works.
What happens to the final chunk if it is shorter than the limit?
By default the splitter keeps the last chunk, even when it is shorter than the target (which you usually want, because it is often a call-to-action or a closing line). If you would rather drop a tiny trailing fragment, turn off Keep last chunk if shorter. The splitter will then discard any final chunk below 25 percent of the target size.
How does character-count splitting handle word boundaries?
Character mode tries to cut at the nearest whitespace before the limit so words are not chopped in half. If a single word is longer than the limit (a URL, a hash, a long compound), the splitter falls back to a hard cut at the exact character count for that chunk only.
Can I split on multiple delimiters at once?
The delimiter field takes one string at a time. For multi-delimiter splitting, run the tool twice: split on the first delimiter, copy all chunks into a single block (or paste them into a doc with a single separator), then split again on the second delimiter.
Why are there fewer chunks than I expected?
Most often this is because your input is shorter than the chunk size, or because consecutive delimiters in the input collapsed empty chunks (the splitter drops empty results). The stat row under the output shows the total chunk count plus the longest and shortest chunk, so you can sanity check the result before pasting.
Is there a limit on input size?
There is no hard cap. The splitter handles inputs of a few megabytes comfortably. Very large inputs (multi-megabyte essays or logs) may take a moment to render every chunk card on screen, but the splitting logic itself is fast.
More wordcounter.ai tools
Other tools you might find useful.