Block: Pulse Block — Zelaxy Docs

The Pulse block extracts text, structure, and data from PDFs, Word documents, PowerPoints, spreadsheets, and images using Pulse. Reach for it when you need to turn uploaded files into clean markdown, HTML, structured output, or chunked text for downstream blocks.

Overview

Property	Value
Type	`pulse`
Category	tools
Color	`#E0E0E0`

When to Use

Convert a PDF, DOCX, PPTX, or XLSX into clean markdown or HTML
Extract text and structure from scanned documents or images
Pull structured output (tables, fields) from a document for downstream processing
Split a large document into chunks for embedding or RAG pipelines
Parse only specific pages of a large file
Retrieve figures, images, or element bounding boxes from a document

Configuration

File Upload

Upload the document to parse directly (basic mode). Accepts PDF, images, DOCX, PPTX, and XLSX up to 50 MB. Required when not using a URL or path.

File URL or Path

Alternative to uploading — provide an HTTP/HTTPS URL or a local path to the file (advanced mode). Required when not uploading a file directly.

Pages

Optional page range to parse, e.g. 1-3,5. Leave blank to parse all pages.

Enable Chunking

Switch to split the extracted output into chunks. When enabled, the chunks output is populated.

Chunk Size

Characters per chunk (default 512) when chunking is enabled. Corresponds to the chunkSize input.

API Key

Your Pulse API key. Required and treated as a password/secret. Use {{PULSE_API_KEY}} to reference an environment variable.

Inputs & Outputs

Inputs:
- fileUpload (json) — Uploaded file object (used in basic mode)
- filePath (string) — URL or path to the file (used in advanced mode)
- pages (string) — Page range to parse (e.g. 1-3,5)
- chunking (boolean) — Whether to split output into chunks
- chunkSize (string) — Characters per chunk
- apiKey (string) — Pulse API key
Outputs:
- markdown (string) — Extracted text in Markdown format
- page_count (number) — Total number of pages
- job_id (string) — Parse job ID
- extraction_url (string) — URL to the extraction result
- html (string) — Extracted content as HTML
- structured_output (json) — Structured extraction result
- chunks (json) — Text split into chunks (if chunking enabled)
- figures (json) — Extracted figures and images
- bounding_boxes (json) — Element bounding box data

Tools

Pulse Document Parser (pulse_parser) — Parses documents (PDFs, images, and Office files) via the Pulse OCR API. Accepts either a file upload object or a URL/path, an optional page range, and optional chunking parameters. Returns markdown, HTML, structured output, chunks, figures, bounding boxes, a job ID, and an extraction URL.

YAML Example

pulse_1:
  type: pulse
  name: "Pulse"
  inputs:
    filePath: "https://example.com/report.pdf"
    pages: "1-3,5"
    chunking: true
    chunkSize: "512"
    apiKey: "{{PULSE_API_KEY}}"
  connections:
    outgoing:
      - target: next-block-id

Pulse Block