AI Utilities workbench
RAG utility

Stop guessing what chunk overlap looks like.

Paste a document, choose chunk size and overlap, and inspect the resulting segments before you build an embedding pipeline.

Static deploymentNo API callsClient-side compute

Document

Word-based local chunking

Chunk map

Exact word units, not fake token precision
R/
Ready to chunk

Paste text and compare fixed-window versus sentence-aware segmentation.

Why words?

Exact units beat fake token counts

Tokenization is model-specific. v0.4.3 uses exact Unicode-separated word units for this visualizer, so the boundaries shown are deterministic. You can translate your target token size into a word target separately if needed.

Copied