核心功能
AI 驱动的视频分析和场景检测自动化音频处理和转录从长视频智能生成剪辑Remotion 驱动的字幕样式设计和叠加FFmpeg 视频导出，具有可配置的质量RevenueCat 订阅管理
开发亮点
适用于 iOS 和 Android 的跨平台 Flutter 开发多阶段 AI 视频处理流水线Remotion 驱动的字幕样式设计FFmpeg 视频导出优化RevenueCat 订阅集成Firebase 认证
技术栈
FlutterDartFirebase AuthRevenueCatRemotionFFmpegAI/ML PipelineCloud Functions
我们的贡献
移动应用程序
开发了适用于 iOS 和 Android 的 Flutter 移动应用程序，提供了视频上传、处理和剪辑生成的流畅用户体验。
AI 处理流水线
开发了多阶段 AI 流水线——视频分析、音频处理、场景处理、剪辑生成——并为每个阶段提供了实时进度跟踪。
字幕系统
集成了 Remotion 用于字幕样式设计，提供多个模板和叠加选项，允许用户自定义生成剪辑上的字幕。
视频导出
实现了基于 FFmpeg 的视频导出，具有可配置的分辨率和质量设置，以实现针对社交媒体优化的输出。
订阅集成
集成了 RevenueCat 进行订阅管理，提供多个套餐等级，通过 iOS 和 Android 上的应用内购买实现变现。
认证
实现了 Firebase 认证，用于安全的用户管理和会话处理。
成果
平台: 跨平台 iOS 和 Android 应用AI 流水线: 多阶段视频分析和剪辑生成变现: RevenueCat 订阅管理导出质量: FFmpeg 优化的社交媒体输出
项目特色
通过一个 AI 驱动的流水线，将长视频转换为可用于社交媒体的短片，该流水线处理视频分析、场景检测、剪辑生成和字幕样式设计——所有操作均通过移动设备完成。
常见问题
MicrocosmWorks built a multi-stage AI pipeline in ReelSnap that first analyzes the full video content, then processes audio for transcription and engagement scoring, detects scene changes and visual highlights, and finally generates optimized short clips at moments of peak interest. The pipeline uses ML models trained on social media engagement patterns to identify segments most likely to perform well as standalone short-form content.
MicrocosmWorks integrated Remotion for caption rendering in ReelSnap, providing multiple template styles with customizable fonts, colors, animations, and positioning. Users can preview styled captions overlaid on their clips before export, with frame-accurate synchronization to the transcribed audio. The Remotion compositions handle word-level timing so captions highlight each word as it is spoken, matching the popular styles used by top social media creators.
MicrocosmWorks implemented FFmpeg-based video export in ReelSnap with configurable resolution, bitrate, and codec settings optimized for each target platform. The export pipeline produces files that meet Instagram Reels, TikTok, and YouTube Shorts specifications without exceeding file size limits. Users can choose between quality presets that balance visual fidelity against upload speed on mobile networks.
MicrocosmWorks integrated RevenueCat for subscription management across both iOS and Android, handling in-app purchases, plan tiers, and usage tracking through a unified SDK. Free users can process a limited number of clips per month, while paid subscribers get higher processing quotas, premium caption styles, and priority rendering in the cloud pipeline. RevenueCat handles receipt validation, subscription restoration, and analytics across both app stores.
MicrocosmWorks built ReelSnap using Flutter for cross-platform development targeting both iOS and Android from a single Dart codebase. Firebase handles authentication and user session management, while the AI processing pipeline runs on cloud infrastructure to keep the mobile app lightweight. At MicrocosmWorks development rates of $20-$40/hr, Flutter-based development typically costs 35-40% less than building separate native apps.