Base 0 as generative origin
In Plexadi's architecture, Base 0 is not an empty binary zero. It is a sealed state in which information is neutralised until an authorised engine brings it back.
技术组合
以新的方式表示信息、在系统之间传输信息,并在目的端重建重要内容。
探索企业试点技术概念的公开概述。技术披露、权利要求和实施细节仅在保密讨论中提供。
Video explainer · 2:54
Why legacy codecs reach their limits with 8K, volumetric video and AR — and how reversible symbolic structures offer a different path.
Concept explainer. Performance statements are inventor-reported and not independently verified.
深入了解架构
从运作角度介绍各个支柱——信息如何流动、与传统系统相比有何变化,以及适用场景。权利要求、公式和实施细节保持保密。
P4 在普通网络不可用的通道上传输精确字节。P4-S 更进一步:它传递含义和引用,使接收方能找到所需资源并在本地组装结果。
发送端表达目的端所需的内容:实体、其关系以及所依赖的资源。
传输语义引用,而非整个成品对象。需要精确字节的内容可使用 P4 以二进制为主、视觉方式为后备的路径。
接收端识别其本地已有或可从可信集合中检索的资源。
结果在目的端组装,包括嵌套依赖项,无需生成式猜测。
传统交付
P4 / P4-S 方法
适用场景
正在申请专利的技术方向的概念性描述。收益属于潜在收益,需要实施测试和独立验证;不声称已获专利授权。
P9 · Base 0 · Concept overview
A high-level view of the mathematical foundation behind P9 and how it connects storage, security and reconstruction across the P-series.
In Plexadi's architecture, Base 0 is not an empty binary zero. It is a sealed state in which information is neutralised until an authorised engine brings it back.
Instead of moving full payloads across vulnerable networks, a compact deterministic seed is sent and the receiver's local base-12 engine rebuilds the content where and when it is needed.
A linear parser or AI scanner that tries to flatten the multi-dimensional structure is designed to see only uniform, high-entropy noise rather than readable data.
Colour and visual data are mapped through percentage-based RGB aligned with duodecimal structures, avoiding the overhead of 24-bit hexadecimal colour spaces.
Concept description only. Protocol sequences, codebook and implementation details remain confidential, and security and reconstruction properties are inventor-reported design goals, not independently verified results.
P7 · B12PPP · Concept overview
How B12PPP aims to remove data at rest as a target and make traffic in transit hard to intercept, replay or manipulate.
Files are not kept on disk as readable static objects. They are distilled into compact mathematical seeds, so a breached disk or cloud bucket is designed to hold no raw data to steal or encrypt for ransom.
B12PPP is designed to remove the Key Management Server as a single point of failure. Key coordinates are masked inside ordinary-looking semantic traffic instead of being held in one attractive target.
Content is rebuilt locally in memory only when an authorised process requests it, and returns to its sealed state when it is closed.
Unauthorised probing of the sealed geometry is designed to drive the attacking process into deterministic failure states rather than towards readable data.
The semantic language used between endpoints rotates automatically, by default every 60 seconds and adjustable to the customer's policy, so captured traffic quickly loses context.
Handshakes and files are split into 40 to 90 parts of unequal percentage sizes that together make up 100%. No single part is intended to carry enough information to reconstruct the whole.
An attacker who captures a single handshake part may hold only around 1% of the whole. Because every part has a dynamic size, there is no fixed boundary showing where one handshake ends and the next begins. To know which parts belong together, an attacker would also need to convert each part back through the correct bases in the correct order, which the design aims to make close to impossible.
Endpoints monitor timing. An unexpected delay of around 4 ms is treated as a possible interception, and the session rotates to an earlier semantic language or is terminated.
Static binary files and handshakes pass through several transformation layers, including base-12 and symbolic glyph representations, before travelling across the network as base-12 strings. The specific layers, bases and their order remain confidential.
Architecture description only. Conversion bases, intermediate encodings, codebooks and algorithms remain confidential and are shared only under NDA. All properties are inventor-reported design goals under development, not independently verified security results.
Plexenc 研究 · 测试报告 v1.2 · 2026 年 8 月 10 日
Plexadi 提供的报告对比了一段合成 8K UHD 片段与一段真实的 NASA“Moving Water”8K DCI 片段。这些是报告所述结果,并非独立复现的测试或 NASA 的背书。
7680 × 4320 · 60 帧/秒 · 180 帧 · 3.00 秒
489 B
报告所述语义流,视频 + 音频
H.264 参照使用 x264 的 medium 预设生成。
8192 × 4096 · 23.976 帧/秒 · 71 帧 · 2.96 秒
445 B
报告所述语义流,视频 + 音频
H.264 参照为原始 NASA 来源的 MP4,未经重新编码。
报告称,两项测试均逐字节完整恢复了所传输的语义描述。报告明确不声称能对原始媒体进行逐像素完美或逐字节一致的恢复。视频和音频重建依赖 AI 接收端;这些字节数并未证明源图像质量。
原始视频大小是等效的未压缩 RGB 占用空间,并非 Plexenc 传输的文件。语义流大小不包括 AI 接收端和其他重建资源。这些结果涉及 Plexenc 语义媒体研究,并非 P4-二进制-视觉传输的确定性原始字节传输能力或 P7-GlyphLock 安全性的证明。
PDF · 4 页 · 源代码和基准测试数据可在保密协议下提供
SE 2630818-9
PRV 收件日期:2026-11-11 · 51已提交的权利要求
SE 2630817-1
PRV 收件日期:2026-11-10 · 11已提交的权利要求
PRV 收件日期:2026-11-11 · 6已提交的权利要求
依据所提供的 PRV 收件确认书确认。收件确认仅证明已提交申请,不代表专利授权或更早的优先权日。
精选架构 · P4-S
P4-S 探索传递含义和引用,而非仅发送整个成品对象,以便接收系统检索相关资源并组装结果。
仅直接交付载荷,并不能组织重建复杂数字实体所需的外部资源、上下文和依赖项。
利用语义信息识别相关资源,并支持跨本地或分布式环境的重建。
资源复用、上下文感知组装,以及无需重新设计整个架构即可演进的资源集合。
目的端所需内容
在目的端组装
媒体资产及其依赖项可能分散在不同系统中。
接收系统可以从可寻址资源中组装资产,而非将每次交付视为孤立的载荷。
P4-S 是一种独立架构,并非仅限十二进制的系统。其描述独立于任何特定数值表示、几何形式、解释技术或存储介质。这些示例说明的是可能的应用,而非已部署产品或实测性能。
P4 是较早的确定性二进制-视觉传输研究。P4-S 是为弥补该早期研究的不足而开展的较新研究,同时提出独立的语义路由与重建架构。这一开发历程并不意味着其在法律上构成 P4 的正式延续或替代。
相关技术方向
Plexadi 技术组合中的其他概念面向数字系统底层的基础设施。它们是独立的技术方向,并非 P4-S 的预设依赖项。
P4 · 确定性传输
P5 · 容器与同步
P6 · 存储与完整性
P7 · 通信安全
P8-BC12 / BASE C12 · 计算
P9 · 基础
安全交换
技术问答
仅依据 Plexadi 已发布页面作答的 AI 助手。任何保密内容均仅在保密协议流程中讨论。
AI 回答可能不完整。回答仅基于公开信息,不构成技术、法律或投资建议。
面向 CTO、CISO、企业架构师和战略合作伙伴。请告诉我们您的身份以及希望评估的内容。详细材料将在双方签署保密协议后共享,专利文件本身不会公开。
讨论用例、集成要求以及您的环境所需的验证。可在双向保密协议下考虑提供详细材料。
研究结果引自所提供的报告;不声称已独立验证或已获专利授权。