AnyFree Text Chunker for AI & RAG
Split long text into token-limited chunks with overlap, for feeding into an LLM context window or embedding pipeline.
How to use this tool
Paste your document, pick a model to determine token size, and set a max tokens per chunk and overlap. The text is split into numbered chunks, each shown with its token count and a copy button.
Useful for preparing documents for retrieval-augmented generation (RAG), embedding APIs, or splitting long input to stay under a model's context limit.
FAQ
- Is my text uploaded anywhere?
- No. Chunking happens entirely in your browser. Your text is never sent to a server or any AI provider.
- What is chunk overlap for?
- Overlap repeats a few tokens from the end of one chunk at the start of the next, so context isn't lost at chunk boundaries. This is common practice for RAG pipelines and embedding search.
- How do I pick a max tokens per chunk value?
- Keep it comfortably under your model's context window or your embedding model's input limit. Common choices for RAG are 300–800 tokens per chunk, depending on how much context each chunk needs to stand alone.
- Does it try to avoid cutting words in half?
- Yes, chunk boundaries snap to the nearest whitespace where possible, so chunks don't end mid-word.