Projects / 專案總覽
這一頁整理目前對外展示的核心專案。每一個專案都先提供站內詳頁，說明方法、適用場景與治理定位，再連接到外部 demo。
This page gathers the core public-facing projects. Each project first provides an internal detail page before routing to its external demo.
基礎展示版，展示語意污染、假中立、責任鏈缺失與 SCBKR 結構審計，作為後續系統的核心骨架。
A foundation showcase covering semantic contamination, false neutrality, responsibility gaps, and SCBKR structural review as the base architecture for later systems.
將語意責任鏈直接用在反詐騙場景，判斷訊息是否具備進入決策鏈的資格，協助建立不輕信的治理流程。
Applies semantic responsibility-chain analysis to anti-scam use cases, asking whether a message deserves entry into a human decision chain.
掛在 LLM 前面的治理展示包，呈現 safety locks、legal rules、auditable logs、trial-audit / proxy-chat 與工程 API。
A governance showcase package in front of LLMs, presenting safety locks, legal rules, auditable logs, proxy-chat, and engineering API concepts.
以授權邊界、失效條件、責任鏈與語意責任雜湊值為核心的治理引擎示範。它不是只問內容能不能用，而是判定一次使用事件在何種邊界內成立、何時失效、以及失效後責任如何重落。
A governance-engine demo centered on licensing boundaries, failure conditions, responsibility chains, and semantic responsibility hash generation. It does not only ask whether content can be used, but determines under which boundary a usage event is valid, when it fails, and how responsibility is reassigned after failure.
以 SCBKR 五維審計、R-Lock 責任鎖與 WHO/WHY/TRUE 閘門，檢查訊息是否具備進入決策鏈的資格。
Uses SCBKR, R-Lock, and WHO/WHY/TRUE gates to verify whether a message is qualified to enter a decision chain.
將網站、影像、金融三類輸入物件編譯為責任鏈結構，並在進入人類決策鏈前完成資格審計。
Compiles web, image, and financial inputs into responsibility-chain structures and audits decision eligibility before human decision-chain entry.
包含 法律責任判決引擎 與 責任敘事成本審計引擎。
用案例展示主體、邊界、後果、責任與成本如何被拆解。點擊進入完整演示。
Includes the Legal Responsibility Verdict Engine and the Responsibility Narrative Cost Audit Engine.
A case-based demo showing how subject, boundary, consequence, accountability, and cost are decomposed. Click to view the full demo.
一個結合新聞播報、責任審計、數位館藏與文明回放的公共界面。
A public interface combining news reporting, responsibility audit, digital archives, and civilizational replay.
這是一個針對長期 AI 記憶組織的開放式索引框架，協助你把記憶做成可查、可重播、可稽核、可路由的結構。
This is an open indexing framework for long-term AI memory organization, designed for replayable, auditable, and routable memory structure.
完整定位：本專案提供「公開可擴充索引層」，幫助你整理跨會話與跨來源記憶，並且可以接你自己的模型；不包含封閉治理與最終判斷核心。
Full positioning: this project ships an open, extensible indexing layer for cross-session memory organization and routing with your own models, while excluding the closed governance/judgment core.
Semantic Firewall / Assistive Interface / Responsibility Chain Demo
這是一個純前端固定範本展示版，不開放自由輸入、不診斷、不治療、不收集個資。它展示「病症／狀態入口 → 通道缺口 → 系統判定 → 一步輔助 → 責任邊界 → Replay 回放鏈」的完整流程。核心理念是：人不是病名，病名只是入口，通道才是補齊點，責任鏈才是穩定器。
A pure frontend controlled preset demo. It does not accept free-text input, diagnose, treat, or collect personal data. It demonstrates the flow: “condition entry → channel gap → system mapping → one-step support → responsibility boundary → replay chain.” Core idea: a person is not a label; a label is only an entry; the channel is what we bridge; responsibility is what stabilizes the process.
許文耀 / 沈耀888π ｜ Wen-Yao Hsu / Shen Yao 888π
語意防火牆創辦人｜Founder of Semantic Firewall
台灣台中｜Taichung, Taiwan
WIF-Pay / FinTech Governance / Payment Responsibility / Hash Replay / Semantic Firewall
一個純前端金融責任鏈示範，將付款事件轉換為可追溯、可回放、可定責的 Hash Replay 責任鏈。示範銀行、付款平台、服務商與客服節點如何被鎖定與審計。不接真實金流，不保存卡號、CVV 或密碼。
A front-end financial responsibility-chain demo that converts payment events into traceable, replayable, and accountable Hash Replay records. It demonstrates how bank, payment platform, service provider, and support nodes can be tracked and audited. No real payment, card data, CVV, or passwords are collected.
付款不是單點交易，付款是一條責任鏈。
A payment is not a single transaction. It is a responsibility chain.
許文耀／沈耀888π ｜ Wen-Yao Hsu / Shen Yao 888π
Founder of Semantic Firewall
Taiwan, Taichung