#
LeapAI CXR — API Documentation
Welcome to the LeapAI CXR API documentation. This guide covers everything you need to integrate the LeapAI Chest X-Ray (CXR) AI analysis engine into your application.
#
What is LeapAI CXR?
LeapAI CXR is an AI-powered Chest X-Ray analysis engine developed by Drongo AI Private Limited. It processes DICOM images and provides automated screening for 13 pathological conditions, including:
#
Key Features
- JWT-based authentication — Secure token-based access for all inference endpoints
- DICOM input support — Accepts standard
.dcm/.dicomfiles - Single inference — Analyse one CXR image at a time with full output artifacts
- Bulk inference — Process an entire folder of DICOM images in a single request
- AI Heatmap generation — Visual overlay highlighting regions of concern
- PDF report generation — Auto-generated clinical-style PDF report per patient
- JSON predictions — Machine-readable prediction output for downstream integration
#
API Endpoint Summary
The server runs on port 8500 by default. Base URL: http://<server-ip>:8500
#
Architecture Overview
┌────────────────┐ ┌──────────────────┐
│ Your App │ POST /single_inference │ LeapAI CXR │
│ (3rd Party) │ ───────────────────────► │ Server (:8500) │
│ │ Authorization: <token> │ │
│ Pre-issued │ │ ┌────────────┐ │
│ JWT Token │ POST /bulk_inference │ │ AI Models │ │
│ │ ───────────────────────► │ │ │ │
│ │ Authorization: <token> │ │ • TB │ │
│ │ │ │ • Cardiac │ │
│ │ ◄─── JSON Response ──── │ │ • Pleural │ │
│ │ │ │ • Opacity │ │
└────────────────┘ │ │ • Segment │ │
│ └────────────┘ │
│ │
│ Output: │
│ • JSON results │
│ • Heatmap PNG │
│ • PDF Report │
└──────────────────┘
#
Getting Started
- Authentication — How to use your pre-issued JWT token
- Single Inference — Run AI analysis on one DICOM image
- Bulk Inference — Process multiple DICOM images at once
- Response Format — Understand the API response structure & error codes
- Python Integration — Ready-to-use Python script for integration