Skip to main content
Use the Pulse Platform when you want to understand a document workflow visually before you write production code. The Playground lets you upload a sample file, add pipeline steps, inspect results, save reusable presets, and generate SDK code from the exact configuration you tested.

Open the Platform

Sign in, create or select an organization, and open the Playground.

Prefer code first?

Use the API quickstart for Python, TypeScript, and REST examples.

​
Create Or Sign In

Start by signing into the Platform, or create an account if your team is new to Pulse. Pulse sign in screen Pulse create account screen

​
What You See First

The first screen is the Playground: upload on the left, extraction settings on the right, and pipeline controls below the configured step. Pulse Platform upload screen with Extract settings open

​
First Pipeline

1

Start a new extraction

Open New Extraction from the sidebar and upload a PDF, spreadsheet, Word document, image, or supported file URL.
2

Keep the first run simple

Run the default Extract step first. This gives you markdown, detected tables, figures, citations, page count, and an extraction_id you can reuse.
3

Inspect the output

Review the Markdown and Tables tabs. For PDFs and images, use the document viewer to compare extracted content against the original page.
4

Add one next step

Add Schema when you need structured JSON, Tables when the output is table-first, or Split when different sections need different treatment.
5

Save and export

Save step presets once the output looks right, then use Show Code to generate Python, TypeScript, or cURL that reproduces your pipeline.

​
Add And Configure Steps

After Extract, add the next processing step from the pipeline menu. Pipeline step menu with Split, Tables, and Schema options As the pipeline grows, each step gets its own configuration panel. Keep the first version narrow, then iterate with reruns and saved presets. Pulse pipeline configured with Extract, Split, and Schema steps

​
Use AI Helpers For Drafts

Schema and Split both include AI helpers. Use them to draft a first schema or topic list from a prompt, then tighten field names, descriptions, required fields, and topic boundaries before saving presets. AI Schema Helper modal with document, schema, and prompt context options AI Split Helper modal with document, split input, and prompt context options

​
What To Try First

GoalStart withAdd next
Read or index a documentExtractChunking or HTML output
Pull invoice or form fieldsExtractSchema
Rebuild financial tablesExtractTables with merge enabled
Route a long document by topicExtractSplit
Extract different fields per sectionExtractSplit, then Schema
Move a tested workflow into codeAny tested pipelineShow Code and presets
If you are unsure which step to add, see Choose Your Workflow. It is the fastest way to decide between Schema, Tables, and Split.

​
Production Handoff

The Platform is not a separate product path from the API. It is the safest way to design the workflow:
  1. Build the pipeline on representative documents.
  2. Save any extract, split, schema, or table presets you want to reuse.
  3. Click Show Code to copy the equivalent SDK call sequence.
  4. Replace the sample file and API key in your app.
  5. Use async processing, webhooks, or batch endpoints when you move to higher volume.
When the pipeline is ready, Show Code gives you the matching Python, TypeScript, or cURL implementation. Show Code modal with generated Python for a Pulse extraction pipeline

​
Next Steps

Choose Your Workflow

Decide which pipeline shape fits your document.

Platform Walkthrough

Learn each part of the Playground in more depth.

API Quickstart

Run the same flow from code.

From Platform to Production

Turn a tested pipeline into production SDK calls.