Pannotate 绘动
Spatially guided AI video creation
基于空间标注的 AI 视频生成工具
Project Website | 项目网站
Website: pannotate.art
Pannotate also has an early web prototype that demonstrates the core product concept and workflow.
Pannotate 已经有一个早期网页原型,用于展示产品概念和基础工作流。
Overview | 项目简介
Pannotate is an AI video creation tool designed around visual guidance. Instead of relying only on long text prompts, users can start from an image, mark the image directly, describe the desired motion, and use those combined inputs to guide AI video generation.
Pannotate 是一个围绕“视觉引导”设计的 AI 视频生成工具。它不只是依赖长文字提示词,而是让用户从一张图片开始,直接在图片上进行标注,再配合动作描述,引导 AI 更准确地生成视频。
The long-term goal is to make image-to-video generation more precise, controllable, and creator-friendly.
项目的长期目标是让图生视频变得更加精准、可控,并且更适合创作者使用。
Core Concept | 核心理念
Most image-to-video tools ask users to describe everything in text. This can be limiting when the user wants to control a specific object, direction, region, or movement.
大多数图生视频工具都要求用户用文字描述一切。但当用户想控制某个具体对象、方向、区域或动作时,纯文字往往不够直观。
Pannotate explores a different workflow:
Pannotate 探索的是另一种工作流:
Upload or select an image
上传或选择一张图片
Annotate directly on the image
直接在图片上进行标注
Use visual marks such as drawing, circles, text, or directional guidance
使用画笔、圆圈、文字或方向标记表达意图
Add a short motion prompt
添加简短的动作描述
Generate and organize short video clips
生成并整理短视频片段
The intended AI input is not just text. It is a combination of:
AI 接收的输入不只是文字,而是以下信息的组合:
This makes the user’s intent more spatially precise.
这样可以让用户的意图在空间位置上表达得更准确。
Current iOS Prototype | 当前 iOS 原型
The current iOS version is a native SwiftUI prototype built for iPhone first.
当前 iOS 版本是一个基于 SwiftUI 构建的原生 iPhone 优先原型。
It focuses on validating the product flow, interface structure, and local prototype interactions before connecting real AI generation services.
当前阶段主要用于验证产品流程、界面结构和本地交互原型,暂未接入真实 AI 生成服务。
Implemented Features | 已实现功能
Current prototype features include:
当前原型已包含以下功能:
Native SwiftUI iOS interface
原生 SwiftUI iOS 界面
iPhone-first layout
iPhone 优先设计
Projects screen for organizing creative work
Projects 页面,用于管理创作项目
Studio screen for selecting images and preparing generation input
Studio 页面,用于选择图片并准备生成输入
Image selection from the device
从设备中选择图片
Lightweight image adjustment flow
简单的图片位置和缩放调整流程
Motion prompt input
动作描述输入
Local mock video generation flow
本地模拟视频生成流程
Outputs screen for generated clips
Outputs 页面,用于展示生成结果
Sequence screen for arranging clips into a larger video flow
Sequence 页面,用于组织多个片段
Profile and Settings screens
Profile 和 Settings 页面
Light / Dark / System appearance switching
浅色 / 深色 / 跟随系统主题切换
Local project and clip management interactions
本地项目和片段管理交互
TestFlight internal testing setup
TestFlight 内部测试配置
Current Status | 当前状态
Pannotate is currently a working local iOS prototype.
Pannotate 目前是一个可以运行的本地 iOS 原型。
The current version is suitable for:
当前版本适合用于:
testing the native iOS user experience
测试原生 iOS 用户体验
validating the Studio to Outputs workflow
验证 Studio 到 Outputs 的生成流程
testing local mock interactions
测试本地模拟交互
collecting feedback through TestFlight internal testing
通过 TestFlight 内部测试收集反馈
The following features are not yet fully implemented:
以下功能尚未完整实现:
real AI video generation API
真实 AI 视频生成 API
backend services
后端服务
user accounts and authentication
用户账号与登录认证
cloud project storage
云端项目存储
real video playback and export pipeline
真实视频播放与导出流程
production-ready annotation export
可用于生产环境的标注导出流程
Tech Stack | 技术栈
- SwiftUI
- Xcode
- iOS Native Development
- App Store Connect
- TestFlight
- GitHub
Repository Structure | 项目结构
The project currently contains:
当前项目主要包含:
Pannotate-iOS/
Main native iOS app source code
iOS 原生应用主要代码
DesignReference/
Design screenshots, product notes, and reference materials
设计截图、产品说明和参考资料
PROJECT_STATUS.md
Project progress and development status summary
项目进度与开发状态总结
README.md
Project overview and collaboration documentation
项目介绍与协作文档
Development Workflow | 开发协作流程
The project uses GitHub for collaboration.
项目使用 GitHub 进行协作。
Recommended workflow:
推荐流程:
main # stable branch
feature/* # feature branches
Suggested process:
建议开发流程:
Pull the latest version from main
从 main 拉取最新代码
Create a new feature branch
创建新的功能分支
Make changes locally in Xcode
在本地 Xcode 中开发
Commit changes with a clear message
使用清晰的提交信息进行 commit
Push the branch to GitHub
推送分支到 GitHub
Open a Pull Request
创建 Pull Request
Review and merge
Review 后合并
Roadmap | 路线图
Planned next steps include:
后续计划包括:
Improve Studio annotation tools
优化 Studio 标注工具
Export annotation data for AI generation
导出可供 AI 使用的标注数据
Connect a real AI video generation API
接入真实 AI 视频生成 API
Add persistent project storage
增加项目持久化存储
Improve clip editing and sequence management
优化视频片段编辑与序列管理
Add real video preview and export
增加真实视频预览与导出
Expand iPadOS support later
后续扩展 iPadOS 支持
Product Direction | 产品方向
Pannotate aims to become a creative tool for AI video generation where users can communicate intent visually, not just textually.
Pannotate 希望成为一个让用户不仅通过文字、也能通过视觉标注表达创作意图的 AI 视频生成工具。
The key design principle is:
核心设计原则是:
Let users show the AI what they mean, directly on the image.
让用户直接在图片上告诉 AI:他们想让哪里发生什么。
Author | 作者
David Woo
License | 许可证
All rights reserved.
版权所有,保留所有权利。
This project is not open-source. Unauthorized copying, redistribution, commercial use, or reuse of the source code is not permitted without written permission.
本项目暂不开源。未经书面许可,不允许复制、再分发、商用或复用本项目源代码。