Skip to content

Web UI

Guide to using Doctra's Gradio-based web interface.

Overview

Doctra provides a user-friendly web interface for document processing without writing code.

Launching the UI

Python

from doctra import launch_ui

# Launch web interface
launch_ui()

Command Line

python -m doctra.ui.app

Module Script

python gradio_app.py

The UI opens at: http://127.0.0.1:7860

Interface Tabs

1. Full Parse

Complete document processing:

  • Upload PDF
  • Configure settings
  • View results
  • Download outputs

2. Tables & Charts

Specialized extraction:

  • Extract charts and/or tables
  • Enable VLM processing
  • Configure API keys
  • Download structured data

3. DocRes

Image restoration:

  • Upload images or PDFs
  • Select restoration task
  • Compare before/after
  • Download enhanced files

4. Enhanced Parser

Combined restoration and parsing:

  • Upload PDF
  • Configure restoration
  • Enable VLM
  • Get comprehensive results

Features

  • Drag & Drop: Easy file upload
  • Real-time Progress: See processing status
  • Preview Results: View output in browser
  • Download ZIP: Get all results packaged
  • Configuration: Adjust all settings
  • API Key Management: Secure key input

Configuration Options

Each tab provides settings for:

  • DPI resolution
  • Language selection
  • VLM provider and API key
  • Restoration tasks
  • Output preferences

Sharing the UI

Launch with public URL:

from doctra import build_demo

demo = build_demo()
demo.launch(share=True)

This generates a temporary public URL for sharing.

Use Cases

  • Non-technical Users: No coding required
  • Quick Processing: Fast one-off document processing
  • Experimentation: Try different settings
  • Demonstrations: Show Doctra capabilities
  • Prototyping: Test before integrating

See Also