# AnchorFact

> AnchorFact is an AI-native knowledge base purpose-built for LLM citations. Every article is structured, source-verified, and GEO-optimized. Content is available in multiple formats (Markdown, JSON-LD, plain text, Turtle/RDF) for maximal AI accessibility.

## Knowledge Base (80 articles)

- [AI Agents (2025-2026)](https://anchorfact.org/kb-2026-00018/index.md): 2025 was widely called the "Year of the Agent" across the AI industry, marking a transition from passive conversational AI to autonomous systems capable of multi-step reasoning, tool use, and independent action. Key developments included the emergence of MCP as an industry-standard integration protocol (December 2025, donated to the Linux Foundation's AAIF), the rise of Computer Use agents that directly manipulate desktop/browser interfaces, and the maturation of multi-agent orchestration frameworks. By 2026, AI agents are deployed across software engineering (Devin, $10.2B valuation), general task automation (Manus, acquired by Meta), and enterprise workflow automation (n8n, Dify, Coze).
- [Attention Mechanism](https://anchorfact.org/kb-2026-00002/index.md): The attention mechanism allows neural networks to dynamically focus on the most relevant parts of input data when producing each output. Introduced by Bahdanau et al. (2014) for machine translation, it solves the information bottleneck of fixed-length context vectors in encoder-decoder architectures by letting the decoder "look back" at all encoder hidden states with learned importance weights.
- [BERT (Bidirectional Encoder Representations from Transformers)](https://anchorfact.org/kb-2026-00003/index.md): BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained language model introduced by Google in 2018 that revolutionized NLP by reading text bidirectionally — considering both left and right context simultaneously. It achieved state-of-the-art results on 11 NLP benchmarks at launch, including GLUE (80.5% → 82.1% absolute improvement), SQuAD v1.1, and SWAG.
- [Constitutional AI](https://anchorfact.org/kb-2026-00020/index.md): Constitutional AI (CAI) is Anthropic's method for training AI systems to be helpful and harmless using explicit written principles — a "constitution" — rather than implicit human preferences. First published in December 2022 (arXiv:2212.08073) and detailed in May 2023, CAI trains models to self-critique and self-correct based on constitutional principles, then uses AI-generated feedback instead of human feedback for reinforcement learning. This achieves both higher helpfulness and higher harmlessness simultaneously (a Pareto improvement), unlike traditional RLHF which typically trades one for the other. Claude is the first model trained entirely with CAI.
- [Diffusion Models](https://anchorfact.org/kb-2026-00009/index.md): Diffusion models are a class of generative models that create data (images, audio, video) by learning to reverse a gradual noising process. Starting from pure random noise, they iteratively denoise toward a coherent output. Popularized by DDPM (Ho et al., 2020, over 15,000 citations on Google Scholar as of 2026), diffusion models power major image generation systems including Stable Diffusion, DALL·E 3, Midjourney, and Imagen. They have surpassed GANs as the dominant paradigm for high-quality image generation.
- [GPT (Generative Pre-trained Transformer) Model Family](https://anchorfact.org/kb-2026-00004/index.md): The GPT (Generative Pre-trained Transformer) model family, developed by OpenAI, represents a lineage of autoregressive language models that progressively scaled from 117M to trillions of parameters. GPT-1 (2018) introduced the "pre-train then fine-tune" paradigm using the Transformer decoder; GPT-2 (2019) demonstrated emergent zero-shot capabilities at 1.5B parameters; GPT-3 (2020) established few-shot learning as a viable paradigm at 175B parameters and was trained on approximately 570GB of filtered text. As of 2026, the GPT family has become a cornerstone of modern AI, with GPT-5.2 reaching trillions of parameters.
- [GraphRAG (Microsoft)](https://anchorfact.org/kb-2026-00019/index.md): GraphRAG is a structured, hierarchical approach to Retrieval-Augmented Generation developed by Microsoft Research, open-sourced on July 2, 2024. Unlike standard RAG which retrieves flat text chunks via vector similarity, GraphRAG constructs a knowledge graph from raw text, detects communities using graph algorithms, and generates hierarchical summaries at multiple levels. This enables both local (entity-level) and global (dataset-level) query understanding. As of May 2026, the open-source repository has 33,157 GitHub stars and is deployed in Microsoft Discovery, an Azure-based agentic platform for scientific research.
- [Large Language Models (LLMs)](https://anchorfact.org/kb-2026-00005/index.md): Large Language Models (LLMs) are deep neural networks trained on massive text corpora to predict and generate human language. They scale from hundreds of millions to trillions of parameters and have demonstrated emergent reasoning, coding, and multi-step planning capabilities that were not explicitly programmed. As of 2026, LLMs power most consumer-facing AI products including ChatGPT, Claude, Gemini, and Grok.
- [LoRA (Low-Rank Adaptation)](https://anchorfact.org/kb-2026-00062/index.md): LoRA (Low-Rank Adaptation) is a parameter-efficient fine-tuning method for large language models, introduced by Hu et al. from Microsoft in 2021 (arXiv:2106.09685, 19,123 citations as of May 2026). Instead of updating all model parameters during fine-tuning, LoRA injects small, trainable low-rank matrices into the model's weight layers, reducing trainable parameters by up to 10,000x while maintaining near full-fine-tuning performance. The method has become the dominant fine-tuning approach in the open-source LLM community, with 13,547 GitHub stars.
- [Model Context Protocol (MCP)](https://anchorfact.org/kb-2026-00016/index.md): The Model Context Protocol (MCP) is an open standard for connecting AI models to external data sources and tools, introduced by Anthropic on November 25, 2024. It replaces the fragmented pattern of custom integrations with a unified client-server protocol. As of December 2025, MCP has been adopted by OpenAI, Google, and Microsoft; hosts over 10,000 public servers; sees 97 million SDK downloads per month; and has been donated to the Linux Foundation under the new Agentic AI Foundation (AAIF).
- [Mixture of Experts (MoE)](https://anchorfact.org/kb-2026-00010/index.md): Mixture of Experts (MoE) is a neural network architecture that divides a model into multiple specialized "expert" sub-networks, with a learned gating mechanism routing each input token to only a subset of experts. This enables models with trillions of total parameters while keeping inference cost proportional to a much smaller "active" parameter count. GPT-4 reportedly uses MoE with 8 experts per layer and approximately 1.76 trillion total parameters but only ~280 billion active per forward pass. MoE is the dominant architecture for frontier-scale models as of 2026.
- [Retrieval-Augmented Generation (RAG)](https://anchorfact.org/kb-2026-00006/index.md): Retrieval-Augmented Generation (RAG) is an AI architecture that combines large language models with external knowledge retrieval to produce more accurate, up-to-date, and source-attributed responses. Instead of relying solely on parametric knowledge (what the model memorized during training), RAG retrieves relevant documents from a knowledge base in real-time and provides them as context to the generation model. Google Scholar (2026) reports over 5,000 citations for the original RAG paper, reflecting its foundational role in modern AI systems.
- [Reinforcement Learning from Human Feedback (RLHF)](https://anchorfact.org/kb-2026-00007/index.md): RLHF (Reinforcement Learning from Human Feedback) is a technique for aligning AI models with human preferences by using human feedback as a reward signal for reinforcement learning. First demonstrated for Atari games and robotics (Christiano et al., 2017), RLHF became the standard alignment method for production LLMs with OpenAI's InstructGPT (2022), which showed that models fine-tuned with RLHF produce more helpful, harmless, and honest outputs than those trained with supervised fine-tuning alone. As of 2026, RLHF is used by virtually all major LLM providers.
- [Transformer Architecture](https://anchorfact.org/kb-2026-00001/index.md): The Transformer is a neural network architecture based solely on attention mechanisms, introduced by Vaswani et al. in 2017. It replaced recurrence and convolution with self-attention, enabling parallel computation and state-of-the-art performance in NLP.
- [C++ Programming Language](https://anchorfact.org/kb-2026-00078/index.md): C++ is a general-purpose programming language created by Bjarne Stroustrup at Bell Labs in 1979 as "C with Classes," first commercially released in 1985. It extends C with object-oriented programming, generic programming (templates), and functional features while maintaining C's performance and memory control. Standardized by ISO (C++23 is current; C++26 in development), C++ powers game engines (Unreal Engine), browsers (Chrome V8), databases (MySQL), operating systems (Windows, macOS), and high-frequency trading systems.
- [C Programming Language](https://anchorfact.org/kb-2026-00077/index.md): C is a general-purpose, procedural programming language created by Dennis Ritchie at Bell Labs between 1969-1973. It is the foundation of modern computing: Unix/Linux kernels, Windows, macOS, embedded systems, and the reference implementations of Python, Ruby, and Node.js are written in C. Standardized by ISO (C23 is the latest, 2024), C remains essential for systems programming, operating systems, and performance-critical applications where direct memory access and minimal runtime overhead are required.
- [CSS (Cascading Style Sheets)](https://anchorfact.org/kb-2026-00069/index.md): CSS (Cascading Style Sheets) is a stylesheet language developed by the W3C that describes the presentation of HTML documents. First proposed by Håkon Wium Lie in 1994, CSS has evolved from basic text styling to a powerful layout and animation engine. Modern CSS (2024+) supports container queries, cascade layers, native nesting, `has()` selector, view transitions, and scroll-driven animations — capabilities that previously required JavaScript.
- [Docker](https://anchorfact.org/kb-2026-00013/index.md): Docker is an open-source platform for developing, shipping, and running applications in lightweight, isolated containers. First released in 2013 by Solomon Hykes at dotCloud, Docker revolutionized software deployment by packaging applications with all their dependencies into portable images that run consistently across any environment. As of 2026, Docker is used by over 15 million developers and is the foundation of modern cloud-native infrastructure.
- [Ethereum](https://anchorfact.org/kb-2026-00068/index.md): Ethereum is a decentralized, open-source blockchain platform with smart contract functionality, proposed by Vitalik Buterin in 2013 and launched in July 2015. It enables developers to build decentralized applications (dApps) on a Turing-complete virtual machine (EVM). As of May 2026, Ethereum is the second-largest cryptocurrency by market capitalization and the dominant platform for DeFi, NFTs, and Web3 applications. Ethereum completed the transition from Proof-of-Work to Proof-of-Stake ("The Merge") in September 2022, reducing energy consumption by approximately 99.95%.
- [Git Version Control System](https://anchorfact.org/kb-2026-00066/index.md): Git is a distributed version control system created by Linus Torvalds in April 2005 for Linux kernel development. It tracks changes to files, enabling multiple developers to collaborate without a central server. Git is the de facto standard for software version control, used by over 90% of developers (Stack Overflow 2024 Survey) and hosting platforms including GitHub, GitLab, and Bitbucket. Its distributed model means every clone is a full repository with complete history.
- [Go Programming Language](https://anchorfact.org/kb-2026-00073/index.md): Go (Golang) is a statically typed, compiled programming language designed at Google by Robert Griesemer, Rob Pike, and Ken Thompson, first released in 2009. It combines the performance of C with the ease of Python — goroutines for concurrency, fast compilation, garbage collection, and a simple syntax. Go powers Docker, Kubernetes, Terraform, Prometheus, and Cloudflare's edge infrastructure. GitHub: 134,013 stars as of May 2026.
- [GraphQL](https://anchorfact.org/kb-2026-00074/index.md): GraphQL is a query language and runtime for APIs developed by Facebook in 2012 and open-sourced in 2015. Unlike REST where the server defines fixed endpoints, GraphQL lets clients request exactly the data they need in a single request — no over-fetching or under-fetching. The specification is maintained by the GraphQL Foundation (part of the Linux Foundation). Key adopters include GitHub, Shopify, Airbnb, and PayPal.
- [Hypertext Transfer Protocol (HTTP)](https://anchorfact.org/kb-2026-00008/index.md): HTTP (Hypertext Transfer Protocol) is the foundational application-layer protocol of the World Wide Web, standardized by the IETF. It defines how clients (browsers, apps) request resources from servers and how servers respond. HTTP has evolved through four major versions: HTTP/0.9 (1991), HTTP/1.0 and 1.1 (1996-1997), HTTP/2 (2015), and HTTP/3 (2022). As of 2026, HTTP/2 and HTTP/3 together carry the majority of global web traffic, and HTTP/3's use of QUIC (UDP-based instead of TCP) represents a fundamental architectural shift.
- [JSON (JavaScript Object Notation)](https://anchorfact.org/kb-2026-00070/index.md): JSON (JavaScript Object Notation) is a lightweight, text-based data interchange format standardized as IETF RFC 8259 and ECMA-404. Derived from JavaScript object literal syntax and popularized by Douglas Crockford in the early 2000s, JSON is the dominant data format for web APIs, configuration files, and data storage. Its simplicity — four primitive types (string, number, boolean, null) plus arrays and objects — has made it the universal lingua franca of data exchange on the web.
- [JWT (JSON Web Token)](https://anchorfact.org/kb-2026-00076/index.md): JWT (JSON Web Token) is a compact, URL-safe token format standardized as IETF RFC 7519. It enables stateless authentication by encoding claims (user identity, permissions, expiry) in a self-contained token signed by the server. JWT is the most widely used token format for web API authentication, replacing server-side sessions in stateless architectures.
- [Kubernetes (K8s)](https://anchorfact.org/kb-2026-00014/index.md): Kubernetes (K8s) is an open-source container orchestration platform originally designed by Google, based on their internal Borg system. Donated to the CNCF in 2015, it automates deployment, scaling, and management of containerized applications across clusters of machines. As of 2026, Kubernetes is the de facto standard for cloud-native infrastructure, used by over 60% of organizations running containers in production according to the CNCF Annual Survey.
- [Linux](https://anchorfact.org/kb-2026-00065/index.md): Linux is an open-source, Unix-like operating system kernel created by Linus Torvalds in 1991. Combined with GNU utilities, it forms the GNU/Linux operating system that powers 100% of the TOP500 supercomputers, 96.3% of the top 1 million web servers, all Android devices (3+ billion), and the majority of cloud infrastructure (AWS, GCP, Azure). Linux is licensed under GPLv2 and maintained by thousands of contributors worldwide, with Linus Torvalds remaining the principal maintainer as of 2026.
- [Markdown](https://anchorfact.org/kb-2026-00071/index.md): Markdown is a lightweight markup language for creating formatted text using a plain-text editor, created by John Gruber in 2004 with significant contributions from Aaron Swartz. Its design goal is readability — Markdown source is intended to be as readable as plain text, unlike HTML's angle brackets. Markdown is the dominant format for documentation (README files, GitHub wikis), static site generators (Jekyll, Hugo, Astro), note-taking apps (Obsidian, Notion), and LLM-optimized content (llms.txt). The CommonMark specification (v0.31.2, 2024) provides a rigorous, unambiguous standard.
- [MongoDB](https://anchorfact.org/kb-2026-00072/index.md): MongoDB is a document-oriented NoSQL database first released in 2009 by MongoDB Inc. (formerly 10gen). Instead of tables and rows, it stores data as JSON-like BSON documents in collections, with a flexible schema. As of May 2026, MongoDB has 28,333 GitHub stars and is one of the most popular NoSQL databases, used by companies including Uber, eBay, and Bosch. MongoDB 8.0 (2024) is the current major release.
- [Next.js](https://anchorfact.org/kb-2026-00060/index.md): Next.js is a React-based full-stack web framework created by Vercel, first released in October 2016. It provides server-side rendering (SSR), static site generation (SSG), API routes, and file-system-based routing out of the box. With 139,553 GitHub stars as of May 2026, it is the most popular React framework and powers websites for Nike, TikTok, Hulu, Twitch, and Notion. The App Router (introduced in Next.js 13, 2022, and stabilized in Next.js 14/15) brought React Server Components as the default rendering model.
- [Node.js](https://anchorfact.org/kb-2026-00059/index.md): Node.js is an open-source, cross-platform JavaScript runtime built on Google's V8 engine, created by Ryan Dahl in 2009. It enables JavaScript to run outside the browser — on servers, desktops, and embedded devices — with an event-driven, non-blocking I/O model. As of May 2026, Node.js has 117,329 GitHub stars and powers the backend of millions of applications including Netflix, PayPal, LinkedIn, and Uber. The current LTS version is Node 24.x (Active LTS since April 2026).
- [OAuth 2.0](https://anchorfact.org/kb-2026-00075/index.md): OAuth 2.0 is an authorization framework standardized as IETF RFC 6749 (2012) that enables third-party applications to obtain limited access to a user's resources without exposing their credentials. It powers virtually all "Sign in with Google/GitHub/Apple" flows and is the foundation of modern API security.
- [PostgreSQL](https://anchorfact.org/kb-2026-00067/index.md): PostgreSQL is an advanced open-source relational database system with over 35 years of active development. First released in 1989 as POSTGRES at UC Berkeley and renamed PostgreSQL in 1996, it emphasizes extensibility, SQL standards compliance, and reliability. As of May 2026, it has 20,982 GitHub stars and is the preferred database for applications requiring ACID compliance, complex queries, and data integrity. PostgreSQL 17 (2024) is the current stable release; PostgreSQL 18 is in development.
- [Python Programming Language](https://anchorfact.org/kb-2026-00012/index.md): Python is a high-level, interpreted programming language created by Guido van Rossum and first released in 1991. It emphasizes code readability through significant whitespace and a clean syntax. As of 2026, Python is the most popular programming language according to the TIOBE Index, used by over 15 million developers worldwide for web development, data science, AI/ML, automation, and scientific computing.
- [React](https://anchorfact.org/kb-2026-00023/index.md): React is an open-source JavaScript library for building user interfaces, developed and maintained by Meta (Facebook). Created by Jordan Walke in 2011 and open-sourced in May 2013, React introduced a component-based architecture and a virtual DOM for efficient UI updates. As of 2026, React is the most widely used frontend library globally, with over 200,000 GitHub stars and billions of weekly NPM downloads. React 19 (released December 2024) introduced Server Components as a stable feature, fundamentally changing the React rendering model.
- [REST API](https://anchorfact.org/kb-2026-00064/index.md): REST (Representational State Transfer) is an architectural style for designing networked applications, defined by Roy Fielding in his 2000 doctoral dissertation at UC Irvine. A REST API uses HTTP methods (GET, POST, PUT, DELETE) to perform CRUD operations on resources identified by URLs, with stateless communication and standardized response formats (typically JSON). REST is the dominant API architecture on the web, used by virtually all major platforms including Google, Amazon, Twitter, and GitHub.
- [Rust Programming Language](https://anchorfact.org/kb-2026-00021/index.md): Rust is a systems programming language focused on safety, concurrency, and performance. Created by Graydon Hoare at Mozilla Research in 2010 and first stabilized in 2015, Rust guarantees memory safety without garbage collection through its ownership and borrowing system. According to the JetBrains State of Rust 2025 survey (February 2026), 26% of Rust developers use it professionally, 65% use it for personal projects, and the language continues to attract new users at an accelerating rate (30% of respondents had started using Rust within the past month). The latest stable version is Rust 1.95.0 (released April 16, 2026). Rust was voted the "most admired" programming language in the Stack Overflow Developer Survey for nine consecutive years (2016-2024).
- [SQL (Structured Query Language)](https://anchorfact.org/kb-2026-00015/index.md): SQL (Structured Query Language) is the standard programming language for managing relational databases. Designed by Donald D. Chamberlin and Raymond F. Boyce at IBM in 1974 and standardized by ANSI in 1986 and ISO in 1987, SQL remains — over 50 years later — the dominant database query language, used by virtually every application that stores structured data. PostgreSQL, MySQL, SQLite, Oracle, and SQL Server collectively power the majority of the world's transactional data systems.
- [Swift Programming Language](https://anchorfact.org/kb-2026-00080/index.md): Swift is a general-purpose, compiled programming language developed by Apple, first announced at WWDC 2014 and open-sourced in December 2015. Designed as a modern replacement for Objective-C, Swift combines safety (optionals, strong typing, automatic memory management via ARC) with performance (LLVM-based compilation). It is the primary language for iOS, macOS, watchOS, and visionOS app development, and is increasingly used for server-side development via the Vapor framework.
- [TCP/IP Protocol Suite](https://anchorfact.org/kb-2026-00011/index.md): TCP/IP is the foundational protocol suite of the Internet, consisting of four abstraction layers that enable reliable, end-to-end communication between devices across heterogeneous networks. TCP (Transmission Control Protocol) provides reliable, ordered, error-checked delivery of byte streams, while IP (Internet Protocol) handles addressing and routing of packets across network boundaries. Standardized in 1981 by Jon Postel through RFC 791 and RFC 793, TCP/IP has scaled from ARPANET to powering over 5 billion Internet users as of 2026.
- [TypeScript](https://anchorfact.org/kb-2026-00022/index.md): TypeScript is a strongly-typed superset of JavaScript developed by Microsoft, first released in October 2012 by Anders Hejlsberg (creator of C# and Turbo Pascal). It adds optional static type checking to JavaScript while compiling to plain JavaScript. TypeScript was the 4th most popular programming language in the 2024 Stack Overflow Developer Survey and is used by over 80% of professional web developers. Its type system enables IDE autocompletion, refactoring, and early error detection — making large-scale JavaScript codebases maintainable.
- [WebAssembly (Wasm) 3.0](https://anchorfact.org/kb-2026-00017/index.md): WebAssembly (Wasm) is a portable binary instruction format for stack-based virtual machines, designed as a compilation target for programming languages. Wasm 3.0, completed on September 17, 2025 (announced by Andreas Rossberg of the W3C WebAssembly Community Group), represents the culmination of 6-8 years of specification work, adding 10 major features including 64-bit addressing, garbage collection, exception handling, tail calls, and multiple memories. It is now available in all major browsers and standalone engines like Wasmtime.
- [WebSocket Protocol](https://anchorfact.org/kb-2026-00063/index.md): WebSocket is a communication protocol providing full-duplex, bidirectional communication over a single TCP connection, standardized as IETF RFC 6455 in 2011. Unlike HTTP's request-response model, WebSocket maintains a persistent connection where both client and server can send data at any time. It is the foundation of real-time web applications including chat systems, live dashboards, collaborative editing, multiplayer games, and financial trading platforms. The protocol has universal browser support and is maintained as a living standard by WHATWG (updated March 2026).
- [YAML (YAML Ain't Markup Language)](https://anchorfact.org/kb-2026-00079/index.md): YAML is a human-readable data serialization format created by Clark Evans in 2001. It uses indentation for structure (like Python) and is the dominant configuration format for DevOps tools (Docker Compose, Kubernetes, Ansible, GitHub Actions, CI/CD pipelines). YAML v1.2.2 (2021) is the current specification.
- [游戏可访问性设计指南](https://anchorfact.org/kb-gd-001/index.md): 游戏可访问性设计指南
- [AI Agent 游戏开发工具](https://anchorfact.org/kb-gd-002/index.md): AI Agent 游戏开发工具
- [AI 对话系统 Prompt 工程指南](https://anchorfact.org/kb-gd-003/index.md): AI 对话系统 Prompt 工程指南
- [注意：Suno官方未提供公开API，以下为社区开源封装项目调用方式](https://anchorfact.org/kb-gd-004/index.md): 1. 概览与分类
- [游戏动画状态机设计](https://anchorfact.org/kb-gd-005/index.md): 游戏动画状态机设计
- [游戏音效设计完全指南](https://anchorfact.org/kb-gd-006/index.md): 游戏音效设计完全指南
- [cloud xr development](https://anchorfact.org/kb-gd-007/index.md): 1. 云游戏技术
- [游戏社区运营](https://anchorfact.org/kb-gd-008/index.md): 游戏社区运营
- [游戏机制与系统设计](https://anchorfact.org/kb-gd-009/index.md): 游戏机制与系统设计
- [游戏 AI 系统设计](https://anchorfact.org/kb-gd-010/index.md): 游戏 AI 系统设计
- [game art pipeline](https://anchorfact.org/kb-gd-011/index.md): 1. 美术管线概览
- [game data analytics](https://anchorfact.org/kb-gd-012/index.md): 2. 核心指标定义与计算
- [游戏设计核心原则](https://anchorfact.org/kb-gd-013/index.md): 游戏设计核心原则
- [游戏开发者社区资源指南](https://anchorfact.org/kb-gd-014/index.md): 游戏开发者社区资源指南
- [游戏引擎对比与技术选型](https://anchorfact.org/kb-gd-015/index.md): 游戏引擎对比与技术选型
- [游戏本地化（Localization / L10n）](https://anchorfact.org/kb-gd-016/index.md): 游戏本地化（Localization / L10n）
- [游戏数学](https://anchorfact.org/kb-gd-017/index.md): 游戏数学
- [概念性 CI 流水线](https://anchorfact.org/kb-gd-018/index.md): 1. 游戏开发生命周期
- [游戏测试方法论](https://anchorfact.org/kb-gd-019/index.md): 游戏测试方法论
- [game ui ux](https://anchorfact.org/kb-gd-020/index.md): 2. 信息设计原则
- [input systems accessibility](https://anchorfact.org/kb-gd-021/index.md): 1. 输入系统架构
- [level design](https://anchorfact.org/kb-gd-022/index.md): 2. 关卡架构类型
- [游戏商业化策略（Monetization Strategy）](https://anchorfact.org/kb-gd-023/index.md): 游戏商业化策略（Monetization Strategy）
- [多人游戏设计（Multiplayer Game Design）](https://anchorfact.org/kb-gd-024/index.md): 多人游戏设计（Multiplayer Game Design）
- [叙事设计](https://anchorfact.org/kb-gd-025/index.md): 叙事设计
- [游戏性能优化](https://anchorfact.org/kb-gd-026/index.md): 游戏性能优化
- [游戏物理系统设计](https://anchorfact.org/kb-gd-027/index.md): 游戏物理系统设计
- [player psychology](https://anchorfact.org/kb-gd-028/index.md): 2. 心流理论
- [程序化内容生成（PCG）](https://anchorfact.org/kb-gd-029/index.md): 程序化内容生成（PCG）
- [游戏渲染管线与着色器](https://anchorfact.org/kb-gd-030/index.md): 游戏渲染管线与着色器
- [save systems design](https://anchorfact.org/kb-gd-031/index.md): 1. 存档系统架构
- [着色器与视觉特效（VFX）](https://anchorfact.org/kb-gd-032/index.md): 着色器与视觉特效（VFX）
- [叙事架构代码实现指南](https://anchorfact.org/kb-gd-033/index.md): 叙事架构代码实现指南
- [剧情架构](https://anchorfact.org/kb-gd-034/index.md): 剧情架构
- [UGC 与 Mod 系统](https://anchorfact.org/kb-gd-035/index.md): UGC 与 Mod 系统
- [Unity Game Engine](https://anchorfact.org/kb-2026-00061/index.md): Unity 是 Unity Technologies 开发的跨平台游戏引擎，首次发布于 2005 年。支持 2D/3D 游戏开发，使用 C# 作为脚本语言，构建目标覆盖 25+ 平台（Windows、macOS、iOS、Android、WebGL、主机等）。截至 2026 年，Unity 是全球使用量最大的游戏引擎，超过 50% 的手游使用 Unity 开发。Unity 2022 LTS 是当前推荐的生产版本，Unity 6（2024年发布）是最新一代。

## API

- [Manifest](https://anchorfact.org/manifest.json): Full article index with metadata
- [JSON-LD endpoint](https://anchorfact.org/kb-2026-00001/index.json): Example structured data (Schema.org TechArticle)
- [Turtle endpoint](https://anchorfact.org/kb-2026-00001/index.ttl): Example RDF knowledge graph data

## Optional

- [GitHub Repository](https://github.com/anchorfact/anchorfact): Source code, governance, and contribution guide