How It Works

The Hybrid Approach

Unlike simple PDF-to-HTML converters that treat forms as static images, our hybrid engine combines two powerful techniques:

1. Phase 1: Programmatic Extraction

   We read the PDF's internal form structure (AcroForm or XFA dictionary) to extract every field's type, name, options, values, validation rules, and position. This is fast, accurate, and costs nothing — no AI needed.

2. Phase 2: AI Vision Refinement

   The extracted fields are assembled into a skeleton HTML form. Then our AI vision model compares the skeleton to the original PDF and refines label text, layout, and styling. Only 1–3 iterations are needed since the structure is already correct.

Supported Field Types

Accessibility Features

• Proper <label> associations for every form field
• <fieldset> and <legend> for radio/checkbox groups
• autocomplete attributes for name, email, phone, address fields
• inputmode for mobile keyboard optimization
• required and aria-required for mandatory fields
• Keyboard navigable with visible focus indicators
• Color contrast meeting WCAG 2.2 AA (4.5:1 ratio)
• Automated axe-core validation with auto-remediation

Output Format

The converted HTML is a standalone file with no external dependencies. It includes:

• Download as JSON — exports all field values as a JSON file
• Download as CSV — exports field values in spreadsheet format
• Print — print-optimized layout with hidden action buttons
• Reset — clears all fields to their default state

Pricing

Form conversion costs 3 credits per page. This is significantly cheaper than our premium vision-only converter (5 credits/page) because the hybrid approach requires fewer AI iterations.