Dzzz

Chunker On-device

Token-sized chunks with overlap, Arabic sentence-aware, JSONL out.

Size 512
Overlap 64
Boundary
Count with
Input 0 chars · 0 words · 0 lines · ≈ 0 tokens
Output

Paste on the left to start.

Paste above to start.

chunks.tsv
1 chunk · 0 tokens — the whole text fits.

Everything runs on your device. Nothing you paste is sent anywhere, and it's cleared when you close the tab.

Token-sized chunks with overlap, Arabic sentence-aware, JSONL out. Everything runs in your browser; nothing you paste is sent anywhere.

By Dzzz AI Updated

About this tool

What you provide

  • Any text.

What you get

  • Token-sized chunks with overlap and a JSONL file.

Limits

  • Claude and Gemini token counts are estimates.

How to use it

  1. Paste text

    Paste or open the document.

  2. Set size and overlap

    Choose chunk size, overlap and boundary.

  3. Export

    Copy the chunks or download JSONL.

Frequently asked questions

How are chunks sized?

By tokens, using the same tokenizers as the token counter, with an optional overlap between neighbouring chunks.

Can I export for a vector store?

Yes. The JSONL tab gives one object per chunk with its id, token count and offsets.