Prompt Chunker
Split long text into token-bounded chunks, with optional overlap between them.
Processed locally in your browser. Nothing you paste is sent anywhere.
- Breaks at paragraph boundaries, so chunks stay coherent instead of cut mid-sentence.
- Optional overlap between consecutive chunks — standard practice in RAG/embedding pipelines, since losing continuity right at a chunk boundary measurably hurts retrieval quality.
- You set the token ceiling per chunk to match your model or embedding limit.
- Runs in your browser — long documents aren't uploaded.
Why this is private
- Processed locally in your browser. Nothing you paste is sent anywhere.
- Open your browser's DevTools → Network tab while using this tool — you won't see a request carrying your input, because it never leaves your device.
- Nothing you enter is uploaded to a server or shared with us — the processing happens entirely on your device.
Overview
Splits a long document into chunks that each stay under a token limit — useful for fitting content into a context window or preparing chunks for embeddings — splitting on paragraph boundaries rather than mid-sentence, with an optional overlap between consecutive chunks so retrieval doesn't lose context right at a cut point.
How to use it
- 1Paste a long document.
- 2Set the maximum tokens per chunk.
- 3Optionally set an overlap (in tokens) so each chunk after the first repeats some trailing context from the one before it.
- 4Copy each chunk individually.
Tips
- Splits happen at paragraph breaks, not mid-sentence — a very long single paragraph that exceeds the limit on its own will still form one (over-limit) chunk rather than being cut arbitrarily. The overlap region is built from whole trailing paragraphs too, for the same reason.
Frequently asked questions
What people use it for
Prepare a document for embeddings with overlap for better retrievalSplit a manual or article into chunks sized for your embedding model's input limit, with overlap so a concept split across a chunk boundary is still retrievable from either side.
Feed a long transcript to a model in partsProcess a recording that won't fit in one context window.
Batch a summarisation jobSummarise each chunk, then summarise the summaries.
Stay within a rate or size limitKeep each request's payload under an API's per-call cap.
Related tools
All ai & llm toolsToken CounterExact token counts for OpenAI models, plus calibrated per-model estimates for Claude, Llama, Gemini and Mistral.Context Window CalculatorSee how much of a model's context window your text uses.AI API Cost CalculatorEstimate monthly LLM API costs from token volume.JSONL ValidatorValidate a JSONL (JSON Lines) file line by line.
Ready to use Prompt Chunker?
Prompt Chunker