Skip to main content

Try Parsefy

The Parsefy Playground lets you test financial document extraction without writing any code or signing up.

Open Playground

Try Parsefy in your browser

How it works

  1. Upload a document: Drag and drop or select a PDF or DOCX file (invoices, receipts, bills)
  2. Define your schema: Use the visual editor or write JSON Schema directly
  3. Extract data: See the structured output with field-level confidence and evidence

Features

No signup required

Start extracting immediately without creating an account

Visual schema builder

Build schemas with a drag-and-drop interface

Field-level confidence

See confidence scores and source evidence for each extracted field

Export code

Get ready-to-use code snippets for your project

Limitations

The playground has usage limits to prevent abuse:
LimitValue
Credits per day10 credits
Max file size10 MB
Rate limit1 request/second
Credits reset at midnight UTC. For unlimited extractions, join the waitlist to get an API key.

From playground to production

Once you’ve tested your schema in the playground, you can export code for your preferred language:
import { Parsefy } from 'parsefy';
import * as z from 'zod';

const client = new Parsefy();

// Schema exported from playground
const schema = z.object({
  // REQUIRED - triggers fallback if below confidence threshold
  invoice_number: z.string().describe('The invoice number'),
  total: z.number().describe('Total amount including tax'),

  // OPTIONAL - won't trigger fallback if missing
  date: z.string().optional().describe('Invoice date'),
  vendor: z.string().optional().describe('Vendor name'),
});

const { object, metadata, verification } = await client.extract({
  file: './invoice.pdf',
  schema,
  confidenceThreshold: 0.85,
  enableVerification: true,
});

console.log(`Confidence: ${object?._meta.confidence_score}`);
if (verification) {
  console.log(`Verification: ${verification.status}`);
}
Remember: All fields are required by default. Mark fields as optional if they might be missing in >20% of your documents to avoid triggering the expensive fallback model.

Next steps