gemini-summary

To generate a zh-TW summary from a PDF or web page, you provide the input to the script, which automatically detects the type and uses Gemini to produce a structured Markdown summary. It extracts web content using Readability and processes PDFs via Gemini APIs.

Yes, you can use Gemini to summarize video and audio files automatically. The Skill detects these media types on demand and leverages the Gemini API to generate a structured zh-TW Markdown summary directly from the provided video or audio input.

Yes, you need a GEMINI_API_KEY and Node.js version 20 or higher to run the summary script. These prerequisites are required to execute the local scripts responsible for preprocessing input and formatting the final zh-TW Markdown output.

The structured Markdown summary output includes sections like 內容摘要, 來源, 核心概述, 重點條列, 關鍵數據, and 行動建議. These predefined sections organize the extracted content into a comprehensive zh-TW overview for better comprehension.

You can override automatic input type detection by using the optional manual type override parameter when running the script. This allows you to explicitly specify whether the input is a URL, PDF, video, or audio file for accurate processing.