Skip to main content

Overview

Narrative API generates AI-powered product thumbnails from videos Key Features:
  • Video processing with dynamic frame sampling (~100 frames)
  • AI-powered frame selection
  • GPU-accelerated video decoding (CUDA)
  • Base64-encoded JPEG output (95% quality)
  • Returns 3 thumbnails for ≤15s videos, 5 for >15s

Narrative OpenAPI Spec

View the OpenAPI specification file

Base URL

Quick Start

Processing Pipeline

  1. Download/receive video
  2. AI selects best frames
  3. Extracted as full-resolution frames
  4. Return base64 JPEGs