Skip to main content
Pulse turns complex documents into reliable text, tables, structured JSON, citations, and reusable extraction artifacts.

Start in the Platform

Build and test your first pipeline visually.

Start with the API

Parse a document from Python, TypeScript, or REST.

Choose a Workflow

Decide between Classify, Extract, Schema, Tables, Split, and Batch.

Security & Compliance

Review data handling, API key, storage, and compliance guidance.
Pulse parses PDFs, spreadsheets, presentations, images, Word documents, and other enterprise files. It is built for document understanding rather than basic OCR: layout, reading order, tables, charts, figures, citations, and structured extraction are first-class parts of the workflow.

​
How it works

Optional first step — Classify: when mixed document types arrive through one intake, route each raw document to the right pipeline before extracting.
  1. Extract a document into markdown, tables, figures, bounding boxes, chunks, and an extraction_id.
  2. Choose a downstream step: Schema for JSON fields, Tables for table-first output, or Split for topic-based page groups.
  3. Chain or scale the workflow with saved presets, async jobs, batch endpoints, webhooks, and SDKs.

​
How to use Pulse

API & SDKs

For developers building automated pipelines. Available in Python, TypeScript, and REST.

Pulse Platform

Visual document processing UI. Upload files, build pipelines, and see results instantly.

​
Pipeline Steps

Classify

Cheaply identify a raw document’s type and route it to the right pipeline — before extracting.

Extract

Parse documents into markdown, tables, and figures with layout-aware processing.

Schema

Apply schemas to extract structured JSON data from documents or topics.

Tables

Reconstruct structured tables, merge cross-page tables, and convert charts.

Split

Divide documents into topic-based page groups for targeted processing.

Sample Documents

Start with realistic PDFs for Platform and API tests.

Bank Statement To JSON

Extract account metadata, balances, transactions, and checks.

Annual Report Split -> Schema

Split a long report and apply per-topic schemas.

Financial Tables

Reconstruct table-first outputs for review and analysis.

Production Webhooks

Handle long-running async jobs cleanly.

​
Get started

Platform Quickstart

Build a pipeline visually.

API Quickstart

Parse your first document in code.

​
Base URL

All API requests should be made to:

​
Need Help?

Support

Troubleshooting, contacts, and support request tips.

Email Support

support@trypulse.ai