普通拍摄被重新定义成带时空锚点的现实输入，而不是又一张待消费的内容卡片。 Casual capture is reframed as reality input with spatiotemporal anchors rather than another content card waiting to be consumed.
// SENSORS + PRIVACY + REALITY INDEX
DISTRIBUTED SENSING
现实如何进入系统 How Reality Enters the System
Oasyce App 的关键不是“用户可以上传照片”，而是它把普通人的随手记录改造成了镜像世界的输入层。每个碎片都先经过隐私过滤，再获得时空锚点、语义标签和索引位置，最终成为 AI 可以调用的现实证据。 Oasyce App matters not because users can upload photos, but because it transforms ordinary casual capture into the input layer of a mirror world. Every fragment is privacy-filtered first, then receives spatial anchors, semantic tags, and index position until it becomes reality evidence callable by AI.
// ORACLE 不是把内容搬上网，而是把现实变成一个可被查询、验证和分润的坐标系。 This is not uploading content to the internet. It is turning reality into a coordinate system that can be queried, verified, and revenue-shared.
隐私门槛前置到设备端，HIGH / CRITICAL 风险在进入系统前就被截断。 The privacy wall moves to the device edge so HIGH / CRITICAL risk is cut off before it enters the system.
碎片最终不是留在相册里，而是进入 4D twin，成为 AI 的现实证据层。 Fragments do not stay trapped in galleries; they move into the 4D twin as the reality-evidence layer for AI.
感知流水线 SENSING PIPELINE
01 CAPTURE
随手记录 Casual capture
gps · timestamp · orientation · altitude 用户不是在“创作内容”，而是在贡献带时空锚点的真实碎片。手机的传感器阵列自动把每次拍摄变成可计算的世界输入。 Users are not “creating content”; they are contributing real fragments with spatiotemporal anchors. The phone’s sensor array automatically turns each capture into computable world input.
02 PRIVACY GATE
设备端隐私关卡 On-device privacy gate
pii detect · face blur · privacy level DataVault 逻辑先在本地运行。敏感信息不出手机，陌生人面部先模糊，HIGH / CRITICAL 风险永不进入后续系统。 DataVault logic runs locally first. Sensitive data never leaves the phone, strangers are blurred first, and HIGH / CRITICAL risk never enters the downstream system.
03 ENRICH
元数据增强 Metadata enrichment
geopoint · dms · capture source · tags 每个碎片都会被补齐来源、位置、时间和轻量语义标签，让它不仅是图片，更是一段可以参与现实索引的证据。 Each fragment is enriched with source, place, time, and lightweight semantic tags so it becomes not merely media, but evidence that can participate in reality indexing.
04 INDEX
建立现实索引 Build the reality index
what + where + when 系统把碎片组织进一个 4D 坐标系。查询“国贸 19:00”不再是搜关键词，而是在现实世界的时空坐标上命中事实。 The system organizes fragments into a 4D coordinate space. Querying “Guomao 19:00” stops being keyword search and becomes a hit against space-time facts in the real world.
05 4DGS FUSION
4DGS 融合 4DGS fusion
fragment -> semantic 3d/4d scene 不同人、不同角度、不同时间的 2D 碎片最终融合成可以被进入、被问询、被再利用的语义场景。 2D fragments from different people, angles, and times ultimately fuse into semantic scenes that can be entered, queried, and reused.
用户获得什么 WHAT THE USER GETS
按需生成的信息流 Generated feed on demand
用户不再滑一个被动推荐流，而是提出问题，让 Agent 从近实时碎片里组织出只对当前时刻有用的答案。 The user no longer scrolls a passive recommendation feed. They ask, and the agent composes an answer useful for this moment from near real-time fragments.
可验证的现实输入 Verifiable reality input
每个答案都可以追溯到来源碎片、地点、时间和贡献者，现实感来自可验证性，而不是文案可信感。 Every answer can be traced back to source fragments, place, time, and contributors. Reality-feel comes from verifiability, not persuasive copy.
贡献自动分润 Automatic contributor payout
一段真实路况、一张排队照片、一次普通路过，都可能成为被调用的数据贡献，并通过协议得到回报。 A real traffic clip, a queue photo, or a routine pass-by can all become callable contributions and be rewarded through the protocol.
镜像世界的输入层 Input layer for the mirror world
这不只是旅行或探店产品，而是世界模型训练、LBS、agent memory 和现实验证的共同入口。 This is not only a travel or discovery product, but a shared entrypoint for world-model training, LBS, agent memory, and reality verification.
生态闭环 ECOSYSTEM HANDOFF
Oasyce App
收集真实碎片 Collect real fragments
DataVault
在设备端做隐私与风险筛选 Run privacy and risk gating on-device
Oasyce Chain
把数据权利、可追溯性和收益分配写进协议 Write rights, traceability, and revenue sharing into protocol state
AI Agent
从现实索引中按需生成答案 Generate answers on demand from the reality index
设计拆解 DESIGN BREAKDOWN
这页核心不是解释技术词，而是证明“镜像世界”并不是 Metaverse 幻觉，而是一条从真实碎片到世界模型输入的产品路线。 The point of this page is not to explain technical terms, but to prove that a mirror world is not metaverse fantasy; it is a product path from real fragments into world-model input.
我把页面拆成 sensing pipeline / generated surfaces / ecosystem handoff，因为这三件事共同定义了 App 的真正产品边界。 The page is split into sensing pipeline / generated surfaces / ecosystem handoff because those three pieces jointly define the real product boundary of the app.
阅读顺序是先理解碎片怎么进入系统，再理解用户如何从系统中获益，最后理解生态如何把价值闭环。 The reading order is: how fragments enter the system, how users benefit from the system, then how the ecosystem closes the value loop.
我保留了镜像世界的系统感，但把信息收束在边框、标签和时空词汇上，避免落回传统“科技感 landing page”的装饰套路。 The page keeps the mirror-world systems feel but concentrates information into borders, labels, and spatiotemporal language instead of falling back to decorative “tech landing page” tropes.
唯一动态仍然是 signal decode，像感知系统上线，而不是把“未来感”堆在画面上。 The only motion remains the signal decode so it feels like a sensing system coming online, rather than piling generic futurism onto the screen.