Sentence-aware workflow
Split text without breaking sentences
Split text without breaking sentences that still fit inside your chosen maximum. Social Threader can prefer complete sentence and paragraph boundaries, then use smaller boundaries only when a sentence is too long to fit.
Quick verdict Turn on Sentences for prose and Paragraphs for multi-section drafts. Every text chunk still obeys the selected maximum, so an over-limit sentence must be divided.
Primary audienceLong-form writers
Core outcomeMore readable chunk boundaries
The problem with raw character cuts
A strict substring can stop halfway through a clause, leave closing punctuation behind, or separate a short lead-in from the sentence it introduces. That makes a thread harder to review and forces manual repair after every cut.
Social Threader first groups words into sentences when sentence-aware mode is enabled. It recognizes the documented abbreviation, decimal, ellipsis, and wrapping-punctuation cases before combining sentences that still fit inside the maximum. Paragraph mode adds another boundary: each paragraph is processed separately.
Important: sentence-aware means boundary-aware, not language understanding. A sentence longer than the maximum cannot remain intact.
How to split text without breaking sentences
- Paste or type the complete draft into the Social Threader editor.
- Select Twitter/X, Bluesky, Threads/Mastodon, or enter a custom maximum.
- Enable Sentences; enable Paragraphs when the draft has multiple sections.
- Optionally enable Enumerate, then review and copy the resulting chunks in order.
The repository’s table-driven tests exercise the public configuration contract. The sentence grouping and over-limit fallback below are copied from the deployed chunking engine:
const sentencesArray = buildSentences(wordsArray, options.breakOnSentences);
/** @type {string[]} */
const baseChunks = [];
let currentChunk = "";
for (const sentence of sentencesArray) {
if (sentence.length > availableLength) {
if (currentChunk.length > 0) {
baseChunks.push(currentChunk);
currentChunk = "";
}
baseChunks.push(...chunkByLength(sentence, availableLength));
continue;
}
const potentialChunk = currentChunk.length > 0 ? `${currentChunk} ${sentence}` : sentence;
if (potentialChunk.length <= availableLength) {
currentChunk = potentialChunk;
} else {
if (currentChunk.length > 0) {
baseChunks.push(currentChunk);
}
currentChunk = sentence;
}
}See the chunking engine on GitHub and its linked repository tests for the complete implementation.
Feature-to-benefit map
| Feature | Why it matters here | Concrete use |
|---|---|---|
| Sentence boundary preference | Keeps complete thoughts together when they fit. | Combine several short sentences into one post without cutting a sentence early. |
| Paragraph-first processing | Prevents unrelated sections from being merged merely because space remains. | Keep an introduction, list, and conclusion as separate thread segments. |
| Maximum-length enforcement | Preserves the actual character constraint after boundary selection. | Divide an unusually long sentence at a usable space or punctuation mark. |
| Live statistics | Makes the source structure visible before splitting. | Confirm that pasted blank lines produced the expected paragraph count. |
Three boundary cases worth reviewing
Abbreviations and decimals
The tests cover tokens such as Dr.,
approx., 5.5, and
p.m. so they do not automatically inflate
the sentence count in the documented cases.
Ellipses and closing quotes
The heuristic considers the next meaningful character and removes wrapping quote or bracket characters before classifying sentence endings.
Lists and paragraph breaks
Paragraph-aware mode recognizes blank-line boundaries and documented list markers, making it useful for agendas or multi-part announcements.
One over-limit sentence
When a sentence itself exceeds the maximum, the engine searches backward for a space or punctuation boundary. A hard cut is the final fallback.
Objections and limitations
Is this a writing or rewriting assistant?
No. Social Threader divides and formats supplied text. The repository does not document summarization, paraphrasing, tone changes, or generated transitions.
Will it preserve every sentence in every language?
No such guarantee is supported. The implementation uses punctuation and abbreviation heuristics documented in the source and tests. Review the output, especially for languages or conventions outside those examples.
What about privacy and data handling?
The chunking engine is browser-side JavaScript in the public repository. The hosted root also loads analytics and a feedback integration, and the repository does not publish a privacy guarantee. Review the source and your own requirements before entering sensitive text.
Frequently asked questions
How do I split text without breaking sentences?
Paste the draft into Social Threader, select a maximum length, and enable Sentences. The tool keeps complete sentences together when they fit within that maximum.
What happens when one sentence is longer than the limit?
The sentence must still be divided. Social Threader looks backward for a space or punctuation boundary and uses a hard character cut only when no usable boundary exists.
Can Social Threader preserve paragraph boundaries too?
Yes. When the input contains multiple paragraphs, enable Paragraphs to process each paragraph separately before applying the selected character limit.
Does sentence-aware mode rewrite my words?
No. The documented engine divides the supplied text and normalizes spacing; it does not summarize, paraphrase, or generate replacement copy.
Does Social Threader post the resulting thread?
No. The web workflow lets you copy each resulting chunk. Posting, scheduling, and account integration remain outside the documented product behavior.
Related Social Threader guides
Keep the readable boundaries that fit
Paste your draft, choose its maximum, and enable sentence or paragraph handling before copying the finished chunks.