Anionex/banana-slides

一个基于nano banana pro🍌的原生AI PPT生成应用,迈向"Vibe PPT"; 支持上传任意模板图片,上传任意素材&智能解析,一句话/大纲/页面描述自动生成PPT,口头修改指定区域、一键导出可编辑ppt - An AI-native slides generator based on nano banana pro🍌

What it solves

Traditional AI presentation tools often suffer from rigid templates, low design quality, and limited flexibility for iterative changes. Banana Slides aims to solve these issues by providing a "Vibe PPT" experience that allows users to create professional, aesthetically consistent presentations without tedious manual formatting.

How it works

The application uses large language models to transform various inputs—such as single ideas, structured outlines, or uploaded documents (PDF, Docx, MD, Txt)—into complete slide decks. It features a natural language editing interface where users can request specific changes (e.g., "change this image to a pie chart") and uses advanced models to ensure high-quality, consistent visual styles across all slides.

Who it’s for

  • Beginners: Users seeking zero-threshold, beautiful presentations without design experience.
  • PPT Professionals: Designers looking for rapid layout and visual element inspiration.
  • Educators: Teachers converting instructional content into illustrated lesson plans.
  • Students: Students needing to quickly complete assignments with a focus on content over formatting.
  • Professionals: Business users needing to quickly visualize proposals or product introductions.

Highlights

  • Flexible Creation: Supports generation from a single prompt, structured outlines, or uploaded files.
  • Natural Language Editing: Allows for conversational, localized, or full-page modifications via text commands.
  • Smart Material Parsing: Automatically extracts text, images, and key points from uploaded documents.
  • Advanced Export Options: Supports exporting to editable PPTX (with animation support), PDF, and MP4 videos with AI-generated voiceovers and subtitles.
  • Visual Control: Includes tools for image editing (overlay/replace/smart erase) and template matching based on reference images.