Convert and format values
Chunk text
Splits long text into pieces for embeddings, knowledge search or an AI step.
Nodes › AIOpen in Neblex
- Good for
- Preparing text for AI
- Role
- Works on the records in a flow
- Find it in
- Nodes › AI
When to use it
- Documents should be loaded into a vector database for knowledge search.
- Text is too long for one model call.
- Pieces should overlap so no sentence loses its context.
Reach for something else when
SplitWhen you are splitting a list, not text.AI taskWhen the AI step can handle long content by itself.
Example
Read the handbook, split it by headings with overlap and load the pieces into a vector database.
How it works
- Settings
- Field [
text]; Split by: Paragraphs [default], Sentences, Headings, Characters or Tokens (approximate); Piece size; Overlap. - Output
- One record per piece:
chunk(the text),index,of(how many pieces) andsource(the field it came from), plus the original record's other fields.
Add it to a flow
- Open the flow from Build › Workflows.
- Open Nodes and find Chunk text under AI, or search for it. Click it or drag it onto the canvas.
- Drag from the previous step's output handle to its input, then open its settings with ✎ Edit.
- Run the flow and click the step to see the records it received and produced.
More in Convert and format values
Hash valueCreates a SHA-256 or SHA-512 fingerprint of a field, or a signature with your key.Nodes › DataFormat dateReformats a date or time into the pattern a system expects.Nodes › DataConvert JSONConverts between records and JSON, CSV or XML text.Nodes › DataScrape websiteExtracts values from HTML with CSS selectors.Nodes › DataRender markdownTurns Markdown into safe HTML.Nodes › DataEncodeEncodes or decodes text (base64, URL, hex, HTML, character sets) or gives each record a stable id.Nodes › Transform