[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tool-sopaco--deepwiki-rs":3,"similar-sopaco--deepwiki-rs":160},{"id":4,"github_repo":5,"name":6,"description_en":7,"description_zh":8,"ai_summary_zh":8,"readme_en":9,"readme_zh":10,"quickstart_zh":11,"use_case_zh":12,"hero_image_url":13,"owner_login":14,"owner_name":15,"owner_avatar_url":16,"owner_bio":17,"owner_company":15,"owner_location":18,"owner_email":19,"owner_twitter":20,"owner_website":21,"owner_url":22,"languages":23,"stars":36,"forks":37,"last_commit_at":38,"license":39,"difficulty_score":40,"env_os":41,"env_gpu":41,"env_ram":41,"env_deps":42,"category_tags":49,"github_topics":52,"view_count":40,"oss_zip_url":20,"oss_zip_packed_at":20,"status":62,"created_at":63,"updated_at":64,"faqs":65,"releases":94},4904,"sopaco\u002Fdeepwiki-rs","deepwiki-rs","Turn code into clarity. Generate accurate technical docs and AI-ready context in minutes—perfectly structured for human teams and intelligent agents.","deepwiki-rs（又名 Litho）是一款基于 Rust 构建的高性能 AI 文档生成引擎，旨在将复杂的源代码自动转化为清晰、专业的架构文档。它能深入分析代码库，在几分钟内生成符合 C4 模型标准的完整技术文档，涵盖上下文图、容器图、组件图及代码级说明，让文档始终与代码变更保持同步。\n\n长期以来，手动维护文档不仅耗时费力，还极易因更新滞后而导致信息失真或结构混乱。deepwiki-rs 彻底解决了这一痛点，它不再依赖人工编写易过时的 Markdown 文件，而是通过 AI 驱动自动生成结构严谨、格式统一且包含可视化图表的\"Repo-Wiki\"，确保团队随时获取准确的架构全貌。\n\n这款工具特别适合开发者、系统架构师及技术负责人使用。无论是需要快速理解遗留代码的新成员，还是希望降低文档维护成本的技术团队，都能从中受益。其独特的技术亮点在于采用 Rust 语言开发，保证了极高的运行性能，同时原生支持为人类团队和智能体（AI Agents）提供结构化的上下文信息，是连接代码现实与架构认知的理想桥梁。","\u003Cp align=\"center\">\n  \u003Cimg height=\"160\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_7cbae956f4a3.webp\">\n\u003C\u002Fp>\n\n\u003Ch3 align=\"center\">Litho (deepwiki-rs)\u003C\u002Fh3>\n\n\u003Cp align=\"center\">\n    \u003Ca href=\".\u002FREADME.md\">English\u003C\u002Fa>\n    |\n    \u003Ca href=\".\u002FREADME_zh.md\">中文\u003C\u002Fa>\n\u003C\u002Fp>\n\u003Cp align=\"center\">💪🏻 High-performance \u003Cstrong>AI-driven\u003C\u002Fstrong> intelligent document generator (DeepWiki-like) built with \u003Cstrong>Rust\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp align=\"center\">📚 Automatically generates high quality \u003Cstrong>Repo-Wiki\u003C\u002Fstrong> for any codebase\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fcrates.io\u002Fcrates\u002Fdeepwiki-rs\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fv\u002Fdeepwiki-rs?color=44a1c9\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fcrates.io\u002Fcrates\u002Fdeepwiki-rs\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fd\u002Fdeepwiki-rs.svg\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Ftree\u002Fmain\u002Fdocs\u002Fen\">\u003Cimg alt=\"Litho Docs\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLitho-Docs-green?logo=Gitbook&color=%23008a60\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Ftree\u002Fmain\u002Fdocs\u002Fzh\">\u003Cimg alt=\"Litho Docs\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLitho-中文-green?logo=Gitbook&color=%23008a60\"\u002F>\u003C\u002Fa>\n  \u003Cimg alt=\"GitHub Actions Workflow Status\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fsopaco\u002Fdeepwiki-rs\u002Frust.yml\">\n\u003C\u002Fp>\n\n\u003Chr \u002F>\n\n# 👋 What's Litho\n\n**Litho** is an AI-powered documentation generation engine that automatically analyzes your source code and generates comprehensive, professional architecture documentation in the C4 model format. No more manual documentation that falls behind code changes - Litho keeps your documentation perfectly in sync with your codebase.\n\nLitho transforms raw code into beautifully structured documentation with context diagrams, container diagrams, component diagrams, and code-level documentation - all automatically generated from your source code.\n\nWhether you're a developer, architect, or technical lead, Litho eliminates the burden of maintaining documentation and ensures your team always has accurate, up-to-date architectural information.\n\n\u003Cp align=\"center\">\n  \u003Cstrong>Transform your codebase into professional architecture documentation in minutes\u003C\u002Fstrong>\n\u003C\u002Fp>\n\n\u003Cdiv style=\"text-align: center; margin: 30px 0;\">\n  \u003Ctable style=\"width: 100%; border-collapse: collapse; margin: 0 auto;\">\n    \u003Ctr>\n      \u003Cth style=\"width: 50%; padding: 15px; background-color: #f8f9fa; border: 1px solid #e9ecef; text-align: center; font-weight: bold; color: #495057;\">Before Litho\u003C\u002Fth>\n      \u003Cth style=\"width: 50%; padding: 15px; background-color: #f8f9fa; border: 1px solid #e9ecef; text-align: center; font-weight: bold; color: #495057;\">After Litho\u003C\u002Fth>\n    \u003C\u002Ftr>\n    \u003Ctr>\n      \u003Ctd style=\"padding: 15px; border: 1px solid #e9ecef; vertical-align: top;\">\n        \u003Cp style=\"font-size: 14px; color: #6c757d; margin-bottom: 10px;\">\u003Cstrong>Manual Documentation\u003C\u002Fstrong>\u003C\u002Fp>\n        \u003Cul style=\"font-size: 13px; color: #6c757d; line-height: 1.6;\">\n          \u003Cli>Outdated, incomplete, or missing documentation\u003C\u002Fli>\n          \u003Cli>Manual updates that fall behind code changes\u003C\u002Fli>\n          \u003Cli>Inconsistent formatting and structure\u003C\u002Fli>\n          \u003Cli>Time-consuming to maintain\u003C\u002Fli>\n          \u003Cli>Hard to navigate and understand\u003C\u002Fli>\n          \u003Cli>Usually just a few markdown files\u003C\u002Fli>\n        \u003C\u002Ful>\n      \u003C\u002Ftd>\n      \u003Ctd style=\"padding: 15px; border: 1px solid #e9ecef; vertical-align: top;\">\n        \u003Cp style=\"font-size: 14px; color: #6c757d; margin-bottom: 10px;\">\u003Cstrong>AI-Generated Documentation\u003C\u002Fstrong>\u003C\u002Fp>\n        \u003Cul style=\"font-size: 13px; color: #6c757d; line-height: 1.6;\">\n          \u003Cli>Automatically generated from codebase\u003C\u002Fli>\n          \u003Cli>Always up-to-date with code changes\u003C\u002Fli>\n          \u003Cli>Professional C4 model structure\u003C\u002Fli>\n          \u003Cli>Consistent formatting and styling\u003C\u002Fli>\n          \u003Cli>Easy to navigate and understand\u003C\u002Fli>\n          \u003Cli>Complete with diagrams, context, and relationships\u003C\u002Fli>\n        \u003C\u002Ful>\n      \u003C\u002Ftd>\n    \u003C\u002Ftr>\n  \u003C\u002Ftable>\n\u003C\u002Fdiv>\n\n\u003Cp align=\"center\">\n  \u003Cstrong>🚀 Litho automatically transforms your messy codebase into beautiful, professional documentation\u003C\u002Fstrong>\n\u003C\u002Fp>\n\n\u003Chr \u002F>\n\n# 😺 Why use Litho\n\n- **Automatically keep documentation in sync** with codebase changes - no more outdated docs\n- **Save hundreds of hours** on manual documentation creation and maintenance\n- **Improve onboarding** for new team members with comprehensive, up-to-date documentation\n- **Enhance code reviews** by providing clear architectural context\n- **Meet compliance requirements** with auditable, automated documentation\n- **Support for multiple programming languages** (Rust, Python, Java, Go, C#, JavaScript, etc.)\n- **Generate professional C4 model diagrams** with context, containers, components, and code\n- **Integrate with CI\u002FCD pipelines** to automatically generate documentation on every commit\n\n🌟 **For:**\n- Development teams of all sizes\n- Open source projects\n- Enterprise software developers\n- Anyone who hates maintaining outdated docs!\n\n❤️ Like **Litho**? Star it 🌟 or [Sponsor Me](https:\u002F\u002Fgithub.com\u002Fsponsors\u002Fsopaco)! ❤️\n\n**Thanks to the kind people**\n\n[![Stargazers repo roster for @sopaco\u002Fdeepwiki-rs](https:\u002F\u002Freporoster.com\u002Fstars\u002Fsopaco\u002Fdeepwiki-rs)](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fstargazers)\n\n# 🌠 Features & Capabilities\n\n### Core Capabilities\n- AI-driven architecture documentation generation from codebase analysis\n- Automatic C4 model diagram creation (Context, Container, Component, Code)\n- Intelligent extraction of code comments, structures, and relationships\n- Multi-language support for various programming languages\n- Customizable template system for documentation output\n\n### Advanced Features\n- **External Knowledge Integration** - Mount external documentation (PDF, Markdown, SQL, etc.) as knowledge sources for enhanced analysis\n- **Database Documentation** - Auto-generate database schema documentation with ERD diagrams for SQL projects\n- Git history analysis for tracking architectural evolution\n- Cross-referencing between code elements and documentation\n- Interactive documentation with embedded diagrams and examples\n- Integration with CI\u002FCD pipelines for automated documentation generation\n\n## 💡 Problem Solved\nLitho solves the common problem of outdated and incomplete technical documentation by automatically generating up-to-date architecture documentation from your source code. No more manual documentation that falls behind code changes - Litho keeps your documentation in sync with your codebase.\n\n# 🌐 Litho Eco Ecosystem\nLitho is part of a broader ecosystem of tools designed to enhance developer productivity and documentation quality. The Litho Eco ecosystem includes complementary tools that work seamlessly with Litho to provide a complete documentation workflow:\n\n## 📘 Litho Book\n**Litho Book** is a high-performance markdown reader built with Rust and Axum, specifically designed to provide an elegant interface for browsing documentation generated by Litho.\n\n### Key Features\n- Real-time markdown rendering with syntax highlighting\n- Full Mermaid chart support for architectural diagrams\n- Intelligent search with fuzzy matching for files and content\n- High-performance architecture with low memory usage\n- AI Intelligent Document Interpretation, Answering Questions\n\n### 🌠 Snapshots\n\u003Cdiv style=\"text-align: center;\">\n  \u003Ctable style=\"width: 100%; margin: 0 auto;\">\n    \u003Ctr>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_ea6525878d02.webp\" alt=\"snapshot-1\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_5ec994c985be.webp\" alt=\"snapshot-2\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n    \u003C\u002Ftr>\n  \u003C\u002Ftable>\n\u003C\u002Fdiv>\n\n### Integration with Litho\nLitho Book serves as the ideal companion application for consuming documentation generated by Litho. The typical workflow is:\n1. Use Litho to generate documentation from your codebase\n2. Use Litho Book to browse and explore the generated documentation with an elegant interface\n\n[Learn more about Litho Book](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Flitho-book)\n\n## 🔧 Mermaid Fixer\n**Mermaid Fixer** is a high-performance AI-driven tool that automatically detects and fixes syntax errors in Mermaid diagrams within Markdown files.\n\n### Key Features\n- Automated scanning of directories for Markdown files\n- Precise detection of Mermaid syntax errors using JS sandbox validation\n- AI-powered intelligent fixing with LLM integration\n- Comprehensive reporting of before\u002Fafter changes\n- Flexible configuration with support for multiple LLM providers\n\n### Integration with Litho\nMermaid Fixer enhances the quality of documentation generated by Litho by automatically fixing syntax errors in Mermaid diagrams. This ensures that all architectural diagrams in your documentation are valid and render correctly.\n\n### 👀 Snapshots\n\u003Cdiv style=\"text-align: center;\">\n  \u003Ctable style=\"width: 100%; margin: 0 auto;\">\n    \u003Ctr>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_ec6d4ce10b4d.webp\" alt=\"snapshot-1\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_9b9f66493fda.webp\" alt=\"snapshot-2\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n    \u003C\u002Ftr>\n  \u003C\u002Ftable>\n\u003C\u002Fdiv>\n\n[Learn more about Mermaid Fixer](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fmermaid-fixer)\n\n## 🤖Agent Skills\nRun in Smithery! [![Run in Smithery](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_cded48b327c8.png)](https:\u002F\u002Fsmithery.ai\u002Fskills?ns=sopaco&utm_source=github&utm_medium=badge)\n\n# 🧠 How it works\n[![zread](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAsk_Zread-_.svg?style=flat&color=00b0aa&labelColor=000000&logo=data%3Aimage%2Fsvg%2Bxml%3Bbase64%2CPHN2ZyB3aWR0aD0iMTYiIGhlaWdodD0iMTYiIHZpZXdCb3g9IjAgMCAxNiAxNiIgZmlsbD0ibm9uZSIgeG1sbnM9Imh0dHA6Ly93d3cudzMub3JnLzIwMDAvc3ZnIj4KPHBhdGggZD0iTTQuOTYxNTYgMS42MDAxSDIuMjQxNTZDMS44ODgxIDEuNjAwMSAxLjYwMTU2IDEuODg2NjQgMS42MDE1NiAyLjI0MDFWNC45NjAxQzEuNjAxNTYgNS4zMTM1NiAxLjg4ODEgNS42MDAxIDIuMjQxNTYgNS42MDAxSDQuOTYxNTZDNS4zMTUwMiA1LjYwMDEgNS42MDE1NiA1LjMxMzU2IDUuNjAxNTYgNC45NjAxVjIuMjQwMUM1LjYwMTU2IDEuODg2NjQgNS4zMTUwMiAxLjYwMDEgNC45NjE1NiAxLjYwMDFaIiBmaWxsPSIjZmZmIi8%2BCjxwYXRoIGQ9Ik00Ljk2MTU2IDEwLjM5OTlIMi4yNDE1NkMxLjg4ODEgMTAuMzk5OSAxLjYwMTU2IDEwLjY4NjQgMS42MDE1NiAxMS4wMzk5VjEzLjc1OTlDMS42MDE1NiAxNC4xMTM0IDEuODg4MSAxNC4zOTk5IDIuMjQxNTYgMTQuMzk5OUg0Ljk2MTU2QzUuMzE1MDIgMTQuMzk5OSA1LjYwMTU2IDE0LjExMzQgNS42MDE1NiAxMy43NTk5VjExLjAzOTlDNS42MDE1NiAxMC42ODY0IDUuMzE1MDIgMTAuMzk5OSA0Ljk2MTU2IDEwLjM5OTlaIiBmaWxsPSIjZmZmIi8%2BCjxwYXRoIGQ9Ik0xMy43NTg0IDEuNjAwMUgxMS4wMzg0QzEwLjY4NSAxLjYwMDEgMTAuMzk4NCAxLjg4NjY0IDEwLjM5ODQgMi4yNDAxVjQuOTYwMUMxMC4zOTg0IDUuMzEzNTYgMTAuNjg1IDUuNjAwMSAxMS4wMzg0IDUuNjAwMUgxMy43NTg0QzE0LjExMTkgNS42MDAxIDE0LjM5ODQgNS4zMTM1NiAxNC4zOTg0IDQuOTYwMVYyLjI0MDFDMTQuMzk4NCAxLjg4NjY0IDE0LjExMTkgMS42MDAxIDEzLjc1ODQgMS42MDAxWiIgZmlsbD0iI2ZmZiIvPgo8cGF0aCBkPSJNNCAxMkwxMiA0TDQgMTJaIiBmaWxsPSIjZmZmIi8%2BCjxwYXRoIGQ9Ik00IDEyTDEyIDQiIHN0cm9rZT0iI2ZmZiIgc3Ryb2tlLXdpZHRoPSIxLjUiIHN0cm9rZS1saW5lY2FwPSJyb3VuZCIvPgo8L3N2Zz4K&logoColor=ffffff)](https:\u002F\u002Fzread.ai\u002Fsopaco\u002Fdeepwiki-rs)\n\n## Four-Stage Processing Pipeline\nLitho's architecture is designed around a four-stage processing pipeline that transforms raw code into comprehensive documentation:\n\n```mermaid\nflowchart TD\n    A[Input: Source Code Repository] --> B[Phase 1: Preprocessing]\n    B --> C[Phase 2: Intelligent Research & Analysis]\n    C --> D[Phase 3: Documentation Generation]\n    D --> E[Phase 4: Verification & Enhancement]\n    E --> F[Output: High-Quality Technical Documentation]\n\n    subgraph Preprocessing Phase\n        B1[Code Scanning & Discovery]\n        B2[Multi-Language Syntax Analysis]\n        B3[Structure & Dependency Extraction]\n        B4[Code Insight Generation]\n        B5[Agent Memory Chunk Initialization]\n        B --> B1 --> B2 --> B3 --> B4 --> B5\n    end\n\n    subgraph Intelligent Research & Analysis Phase\n        C1[System Context Researcher]\n        C2[Domain Module Detector]\n        C3[Workflow Researcher]\n        C4[Boundary Analyzer]\n        C5[Key Module Insight Officer]\n        C6[Agent Memory Chunk Read\u002FWrite]\n        C7[ReAct Reasoning Loop]\n        C --> C1 --> C2 --> C3 --> C4 --> C5 --> C6 --> C7\n    end\n\n    subgraph Documentation Generation Phase\n        D1[Overview Documentation Editor]\n        D2[Architecture Documentation Editor]\n        D3[Workflow Documentation Editor]\n        D4[Boundary Documentation Editor]\n        D5[Key Module Editor]\n        D6[Agent Memory Chunk Reading]\n        D7[High-Quality Documentation Assembly]\n        D --> D1 --> D2 --> D3 --> D4 --> D5 --> D6 --> D7\n    end\n\n    subgraph Verification & Enhancement Phase\n        E1[Mermaid Syntax Verification]\n        E2[Documentation Integrity Check]\n        E3[Diagram Auto-Repair]\n        E4[Quality Report Generation]\n        E5[Final Documentation Output]\n        E --> E1 --> E2 --> E3 --> E4 --> E5\n    end\n\n    style B fill:#e3f2fd,stroke:#1976d2\n    style C fill:#f3e5f5,stroke:#7b1fa2\n    style D fill:#e8f5e8,stroke:#388e3c\n    style E fill:#fff3e0,stroke:#e65100\n```\n\n### Preprocessing Stage\nLitho begins by scanning your entire codebase to identify source files, extract metadata, and analyze project structure. This stage:\n- Discovers all source code files across multiple languages\n- Parses file structures and identifies key components\n- Extracts comments, documentation strings, and code annotations\n- Identifies dependencies between modules and components\n- Builds a comprehensive representation of your codebase\n\n```mermaid\nflowchart TD\nA[Preprocessing Agent] --> B[Structure Extractor]\nA --> C[Original Document Extractor]\nA --> D[Code Analysis Agent]\nA --> E[Relationship Analysis Agent]\nB --> F[Project Structure]\nC --> G[Original Document Materials]\nD --> H[Core Code Insights]\nE --> I[Code Dependencies]\nF --> J[Store to Memory]\nG --> J\nH --> J\nI --> J\n```\n\n### Research Stage\nIn this AI-powered stage, Litho analyzes the code structure to understand the architectural intent:\n- Applies machine learning models to identify patterns and relationships\n- Infers architectural roles from code structure and naming conventions\n- Determines component boundaries and service responsibilities\n- Maps dependencies and data flow between components\n- Identifies potential architectural smells and anti-patterns\n- Generates context-aware documentation for each component\n\n```mermaid\nflowchart TD\nA[Research Orchestrator] --> B[SystemContext Researcher]\nA --> C[Domain Module Detector]\nA --> D[Architecture Researcher]\nA --> E[Workflow Researcher]\nA --> F[Key Module Insights]\nB --> G[System Context Report]\nC --> H[Domain Module Report]\nD --> I[Architecture Analysis Report]\nE --> J[Workflow Analysis Report]\nF --> K[Module Deep Insights]\nG --> Memory\nH --> Memory\nI --> Memory\nJ --> Memory\nK --> Memory\n```\n\n### Composition and Output Stage\nLitho combines the analyzed information into a structured documentation format:\n- Generates C4 model diagrams (Context, Container, Component, Code)\n- Creates hierarchical documentation structure with clear navigation\n- Embeds relevant code examples and explanations\n- Applies consistent styling and formatting across all documentation\n- Adds cross-references between related components and diagrams\n\n```mermaid\nflowchart TD\nA[Document Composer] --> B[Overview Editor]\nA --> C[Architecture Editor]\nA --> D[Module Insight Editor]\nB --> E[Overview Document]\nC --> F[Architecture Document]\nD --> G[Module Documents]\nE --> H[Document Tree]\nF --> H\nG --> H\nH --> I[Disk Outlet]\nI --> J[Output Directory]\n```\n\n### Validation and Enhancement Stage\nThe final stage ensures documentation quality and completeness:\n- Validates diagram syntax and consistency\n- Checks for completeness of documentation coverage\n- Identifies gaps in documentation and suggests improvements\n- Integrates with Mermaid Fixer to ensure all diagrams render correctly\n- Generates statistics and reports on documentation coverage\n- Creates an index and table of contents for easy navigation\n\n# 🏗️ Architecture Overview\n\n**Litho** features a sophisticated modular architecture designed for high performance, extensibility, and intelligent analysis. The system implements a multi-stage workflow with specialized AI agents and comprehensive caching mechanisms.\n\n```mermaid\ngraph LR\n    subgraph Input Phase\n        A[CLI Startup] --> B[Load Configuration]\n        B --> C[Scan Structure]\n        C --> D[Extract README]\n    end\n    subgraph Analysis Phase\n        D --> E[Language Parsing]\n        E --> F[AI-Enhanced Analysis]\n        F --> G[Store in Memory]\n    end\n    subgraph Reasoning Phase\n        G --> H[Orchestrator Startup]\n        H --> I[System Context Analysis]\n        H --> J[Domain Module Detection]\n        H --> K[Workflow Analysis]\n        H --> L[Key Module Insights]\n        I --> M[Store in Memory]\n        J --> M\n        K --> M\n        L --> M\n    end\n    subgraph Orchestration Phase\n        M --> N[Orchestration Hub Startup]\n        N --> O[Generate Project Overview]\n        N --> P[Generate Architecture Diagram]\n        N --> Q[Generate Workflow Documentation]\n        N --> R[Generate Module Insights]\n        O --> S[Write to DocTree]\n        P --> S\n        Q --> S\n        R --> S\n    end\n    subgraph Output Phase\n        S --> T[Persist Documents]\n        T --> U[Generate Summary Report]\n    end\n```\n\n## Core Modules\nLitho's architecture consists of several interconnected modules that work together to deliver seamless documentation generation:\n\n- **Code Scanner**: Discovers and analyzes source code files across multiple languages\n- **Language Parser**: Extracts structural information from code using language-specific parsers\n- **Architecture Analyzer**: AI-powered component that infers architectural patterns and relationships\n- **Diagram Generator**: Creates C4 model diagrams using Mermaid syntax\n- **Documentation Formatter**: Structures content into organized, navigable documentation\n\n## Core Process\nThe core processing flow follows a deterministic pipeline:\n1. **Scan** - Discover and analyze source code files\n2. **Parse** - Extract structural and semantic information\n3. **Analyze** - Apply AI models to infer architecture and relationships\n4. **Generate** - Create diagrams and documentation content\n5. **Format** - Structure content into organized documentation\n6. **Export** - Output in desired format(s)\n\n```mermaid\nsequenceDiagram\nparticipant Main as main.rs\nparticipant Workflow as workflow.rs\nparticipant Context as GeneratorContext\nparticipant Preprocess as PreProcessAgent\nparticipant Research as ResearchOrchestrator\nparticipant Doc as DocumentationOrchestrator\nparticipant Outlet as DiskOutlet\nMain->>Workflow : launch(config)\nWorkflow->>Context : Create context (LLM, Cache, Memory)\nWorkflow->>Preprocess : execute(context)\nPreprocess->>Context : Store project structure and metadata\nContext-->>Workflow : Preprocessing complete\nWorkflow->>Research : execute_research_pipeline(context)\nResearch->>Research : Execute multiple research agents in parallel\nloop Each Research Agent\nResearch->>StepForwardAgent : execute(context)\nStepForwardAgent->>Context : Validate data sources\nStepForwardAgent->>AgentExecutor : Call prompt or extract\nAgentExecutor->>LLMClient : Initiate LLM request\nLLMClient->>CacheManager : Check cache\nalt Cache hit\nCacheManager-->>LLMClient : Return cached result\nelse Cache miss\nLLMClient->>LLM : Call LLM API\nLLM-->>LLMClient : Return raw response\nLLMClient->>CacheManager : Store result to cache\nend\nLLMClient-->>AgentExecutor : Return processed result\nAgentExecutor-->>StepForwardAgent : Return result\nStepForwardAgent->>Context : Store result to Memory\nend\nResearch-->>Workflow : Research complete\nWorkflow->>Doc : execute(context, doc_tree)\nDoc->>Doc : Call multiple composition agents to generate docs\nDoc-->>Workflow : Documentation generation complete\nWorkflow->>Outlet : save(context)\nOutlet-->>Workflow : Storage complete\nWorkflow-->>Main : Process finished\n```\n\n# 🖥 Getting Started\n### Prerequisites\n- [**Rust**](https:\u002F\u002Fwww.rust-lang.org) (version 1.70 or later)\n- [**Cargo**](https:\u002F\u002Fdoc.rust-lang.org\u002Fcargo\u002F)\n\n### Installation\n#### Option 1: Install from crates.io (Recommended)\n```sh\ncargo install deepwiki-rs\n```\n\n#### Option 2: Build from Source\n1. Clone the repository:\n    ```sh\n    git clone https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs.git\n    ```\n2. Navigate to the project directory:\n    ```sh\n    cd deepwiki-rs\n    ```\n3. Build the project:\n    ```sh\n    cargo build --release\n    ```\n4. The compiled binary will be available in the `target\u002Frelease` directory.\n\n# 🚀 Usage\n**Litho** provides a simple command-line interface to generate documentation from your codebase. For more configuration parameters, refer to the [CLI Options Detail](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fblob\u002Fmain\u002Fdocs\u002F5%E3%80%81%E8%BE%B9%E7%95%8C%E8%B0%83%E7%94%A8.md#litho).\n\n### Basic Command\n```sh\ndeepwiki-rs -p .\u002Fmy-project -o .\u002Fdocs\n\n# Generate documentation in the target language.\ndeepwiki-rs --target-language en -p .\u002Fmy-project\n\ndeepwiki-rs --target-language ja -p .\u002Fmy-project\n```\n\nThis command will:\n- Scan all files in `.\u002Fmy-project`\n- Analyze the code structure and relationships\n- Generate comprehensive C4 architecture documentation\n- Save the output to `.\u002Flitho.docs` directory\n\n### Documentation Generation\nLitho supports several options for generating documentation:\n\n```sh\n# Generate documentation with default settings\ndeepwiki-rs skip certain processing stages in the generation workflow\ndeepwiki-rs --skip-preprocessing --skip-research\n```\n\n### Advanced Options\n```sh\n# Turn off ReAct Mode to avoid auto-scanning project files via tool-calls\ndeepwiki-rs -p .\u002Fsrc --disable-preset-tools --llm-api-base-url \u003Cyour llm provider base-api> --llm-api-key \u003Cyour api key> --model-efficient GPT-5-mini\n\n# Set up both the efficient model and the powerful model simultaneously\ndeepwiki-rs -p .\u002Fsrc --model-efficient GPT-5-mini --model-poweruful GPT-5-Pro --llm-api-base-url \u003Cyour llm provider base-api> --llm_api_key \u003Cyour api key> --model-efficient GPT-5-mini\n```\n\n## 📚 External Knowledge Integration\n\nLitho supports mounting external documentation as knowledge sources to enhance generated documentation with business context and architectural decisions.\n\n### Supported Document Types\n- **PDF** - Architecture diagrams, design documents\n- **Markdown** - Technical documentation, ADRs\n- **SQL** - Database schema files\n- **YAML\u002FJSON** - API specifications (OpenAPI), configurations\n- **Text** - Plain text documentation\n\n### Knowledge Categories\nDocuments are organized into categories for targeted delivery to specific agents:\n- `architecture` - System architecture and C4 model docs\n- `database` - Schema, ERD, and data model documentation\n- `api` - API specifications and endpoint docs\n- `deployment` - Infrastructure and DevOps documentation\n- `adr` - Architecture Decision Records\n- `workflow` - Business processes and workflows\n- `general` - Uncategorized general documentation\n\n### Sync Knowledge Command\n```sh\n# Sync external knowledge sources (processes and caches local docs)\ndeepwiki-rs sync-knowledge\n\n# Force sync even if cache is fresh\ndeepwiki-rs sync-knowledge --force\n```\n\n### Configuration Example (litho.toml)\n```toml\n[knowledge.local_docs]\nenabled = true\ncache_dir = \".litho\u002Fcache\u002Fknowledge\u002Flocal_docs\"\nwatch_for_changes = true\n\n# Default chunking for large documents\n[knowledge.local_docs.default_chunking]\nenabled = true\nmax_chunk_size = 8000\nchunk_overlap = 200\nstrategy = \"semantic\"  # Options: semantic, paragraph, fixed\nmin_size_for_chunking = 10000\n\n# Architecture documentation category\n[[knowledge.local_docs.categories]]\nname = \"architecture\"\ndescription = \"System architecture documentation\"\npaths = [\n    \"docs\u002Farchitecture\u002F**\u002F*.md\",\n    \"docs\u002Fdesign\u002F**\u002F*.pdf\"\n]\ntarget_agents = [\n    \"SystemContextResearcher\",\n    \"ArchitectureResearcher\",\n    \"ArchitectureEditor\"\n]\n\n# Database documentation category\n[[knowledge.local_docs.categories]]\nname = \"database\"\ndescription = \"Database schema documentation\"\npaths = [\n    \"docs\u002Fdatabase\u002F**\u002F*.md\",\n    \"docs\u002Fschema\u002F**\u002F*.sql\"\n]\ntarget_agents = [\n    \"ArchitectureResearcher\",\n    \"DomainModulesDetector\",\n    \"KeyModulesInsight\"\n]\n```\n\n## 🗄️ Database Documentation\n\nLitho automatically analyzes SQL database projects (`.sqlproj`) and SQL files to generate comprehensive database documentation including:\n\n- **Database Projects** - SQL Server project structure\n- **Tables** - Schema, columns, data types, constraints, primary keys\n- **Views** - View definitions and referenced tables\n- **Stored Procedures** - Parameters, operations, accessed tables\n- **Functions** - Scalar and table-valued functions\n- **Relationships** - Foreign keys and implicit references (with ERD diagrams)\n- **Data Flows** - ETL operations and data movement patterns\n\n### Database Analysis Features\n```\n📊 Database code distribution: Projects(2) SQL Files(15) DAO(3)\n✅ Database overview analysis completed:\n   - Database projects: 2 items\n   - Tables: 12 items\n   - Views: 5 items\n   - Stored procedures: 8 items\n   - Functions: 3 items\n   - Table relationships: 6 items\n   - Data flows: 4 items\n   - Confidence: 8.5\u002F10\n```\n\n### Generated Database Documentation\nThe database documentation is automatically included in the output as `6.Database-Overview.md` with:\n- Summary statistics table\n- Detailed table schemas with column definitions\n- Mermaid ER diagrams showing relationships\n- Stored procedure documentation\n- Data flow descriptions\n\n## 📁 Output Structure\nLitho generates a well-organized documentation structure:\n\n```\nproject-docs\u002F\n├── 1. Project Overview      # Project overview, core functionality, technology stack\n├── 2. Architecture Overview # Overall architecture, core modules, module breakdown\n├── 3. Workflow Overview     # Overall workflow, core processes\n├── 4. Deep Dive\u002F            # Detailed technical topic implementation documentation\n│   ├── Topic1.md\n│   ├── Topic2.md\n├── 5. Boundary-Interfaces   # API endpoints, external integrations\n├── 6. Database-Overview     # Database schema, tables, relationships (SQL projects only)\n```\n\n# 🤝 Contribute\nWe welcome all forms of contributions! Report bugs or submit feature requests through [GitHub Issues](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fissues).\n\n## Ways to Contribute\n- **Language Support**: Add support for additional programming languages\n- **Template Creation**: Design new documentation templates and styles\n- **Diagram Enhancements**: Improve Mermaid diagram generation algorithms\n- **Performance Optimization**: Enhance processing speed and memory usage\n- **Test Coverage**: Add comprehensive test cases for various code patterns\n- **Documentation**: Improve project documentation and usage guides\n- **Bug Fixes**: Help identify and fix issues in the codebase\n\n## Development Contribution Process\n1. Fork this project\n2. Create a feature branch (`git checkout -b feature\u002Famazing-feature`)\n3. Commit your changes (`git commit -m 'Add some amazing feature'`)\n4. Push to the branch (`git push origin feature\u002Famazing-feature`)\n5. Create a Pull Request\n\n# 🪪 License\n**MIT**. A copy of the license is provided in the [LICENSE](LICENSE) file.\n\n# 👨 About Me\n> 🚀 Help me develop this software better by [sponsoring on GitHub](https:\u002F\u002Fgithub.com\u002Fsponsors\u002Fsopaco)\n\nAn experienced internet veteran, having navigated through the waves of PC internet, mobile internet, and AI applications. Starting from an individual mobile application developer to a professional in the corporate world, I possess rich experience in product design and research and development. Currently, I am employed at [Kuaishou](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FKuaishou), focusing on the R&D of universal front-end systems and AI exploration.\n\nGitHub: [sopaco](https:\u002F\u002Fgithub.com\u002Fsopaco)\n","\u003Cp align=\"center\">\n  \u003Cimg height=\"160\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_7cbae956f4a3.webp\">\n\u003C\u002Fp>\n\n\u003Ch3 align=\"center\">Litho（deepwiki-rs）\u003C\u002Fh3>\n\n\u003Cp align=\"center\">\n    \u003Ca href=\".\u002FREADME.md\">English\u003C\u002Fa>\n    |\n    \u003Ca href=\".\u002FREADME_zh.md\">中文\u003C\u002Fa>\n\u003C\u002Fp>\n\u003Cp align=\"center\">💪🏻 基于\u003Cstrong>Rust\u003C\u002Fstrong>构建的高性能、\u003Cstrong>AI驱动\u003C\u002Fstrong>的智能文档生成器（类似DeepWiki）\u003C\u002Fp>\n\u003Cp align=\"center\">📚 自动为任何代码库生成高质量的\u003Cstrong>Repo-Wiki\u003C\u002Fstrong>\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fcrates.io\u002Fcrates\u002Fdeepwiki-rs\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fv\u002Fdeepwiki-rs?color=44a1c9\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fcrates.io\u002Fcrates\u002Fdeepwiki-rs\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fd\u002Fdeepwiki-rs.svg\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Ftree\u002Fmain\u002Fdocs\u002Fen\">\u003Cimg alt=\"Litho Docs\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLitho-Docs-green?logo=Gitbook&color=%23008a60\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Ftree\u002Fmain\u002Fdocs\u002Fzh\">\u003Cimg alt=\"Litho Docs\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLitho-中文-green?logo=Gitbook&color=%23008a60\"\u002F>\u003C\u002Fa>\n  \u003Cimg alt=\"GitHub Actions Workflow Status\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fsopaco\u002Fdeepwiki-rs\u002Frust.yml\">\n\u003C\u002Fp>\n\n\u003Chr \u002F>\n\n# 👋 什么是Litho\n\n**Litho**是一款基于AI的文档生成引擎，能够自动分析您的源代码，并以C4模型格式生成全面、专业的架构文档。再也不用担心手动维护的文档会落后于代码变更——Litho可以让您的文档始终与代码库保持完美同步。\n\nLitho将原始代码转化为结构清晰、内容丰富的文档，包括上下文图、容器图、组件图以及代码级别的详细说明，所有这些都由您的源代码自动生成。\n\n无论您是开发者、架构师还是技术负责人，Litho都能减轻您维护文档的负担，确保团队始终拥有准确、最新的架构信息。\n\n\u003Cp align=\"center\">\n  \u003Cstrong>在几分钟内将您的代码库转化为专业架构文档\u003C\u002Fstrong>\n\u003C\u002Fp>\n\n\u003Cdiv style=\"text-align: center; margin: 30px 0;\">\n  \u003Ctable style=\"width: 100%; border-collapse: collapse; margin: 0 auto;\">\n    \u003Ctr>\n      \u003Cth style=\"width: 50%; padding: 15px; background-color: #f8f9fa; border: 1px solid #e9ecef; text-align: center; font-weight: bold; color: #495057;\">使用Litho之前\u003C\u002Fth>\n      \u003Cth style=\"width: 50%; padding: 15px; background-color: #f8f9fa; border: 1px solid #e9ecef; text-align: center; font-weight: bold; color: #495057;\">使用Litho之后\u003C\u002Fth>\n    \u003C\u002Ftr>\n    \u003Ctr>\n      \u003Ctd style=\"padding: 15px; border: 1px solid #e9ecef; vertical-align: top;\">\n        \u003Cp style=\"font-size: 14px; color: #6c757d; margin-bottom: 10px;\">\u003Cstrong>手动维护的文档\u003C\u002Fstrong>\u003C\u002Fp>\n        \u003Cul style=\"font-size: 13px; color: #6c757d; line-height: 1.6;\">\n          \u003Cli>文档过时、不完整或缺失\u003C\u002Fli>\n          \u003Cli>手动更新跟不上代码变化\u003C\u002Fli>\n          \u003Cli>格式和结构不统一\u003C\u002Fli>\n          \u003Cli>维护耗时\u003C\u002Fli>\n          \u003Cli>难以浏览和理解\u003C\u002Fli>\n          \u003Cli>通常只有几份Markdown文件\u003C\u002Fli>\n        \u003C\u002Ful>\n      \u003C\u002Ftd>\n      \u003Ctd style=\"padding: 15px; border: 1px solid #e9ecef; vertical-align: top;\">\n        \u003Cp style=\"font-size: 14px; color: #6c757d; margin-bottom: 10px;\">\u003Cstrong>AI生成的文档\u003C\u002Fstrong>\u003C\u002Fp>\n        \u003Cul style=\"font-size: 13px; color: #6c757d; line-height: 1.6;\">\n          \u003Cli>从代码库自动生成\u003C\u002Fli>\n          \u003Cli>始终与代码变化保持同步\u003C\u002Fli>\n          \u003Cli>采用专业的C4模型结构\u003C\u002Fli>\n          \u003Cli>格式和样式统一\u003C\u002Fli>\n          \u003Cli>易于浏览和理解\u003C\u002Fli>\n          \u003Cli>包含完整的图表、上下文关系等\u003C\u002Fli>\n        \u003C\u002Ful>\n      \u003C\u002Ftd>\n    \u003C\u002Ftr>\n  \u003C\u002Ftable>\n\u003C\u002Fdiv>\n\n\u003Cp align=\"center\">\n  \u003Cstrong>🚀 Litho可以自动将凌乱的代码库转化为美观、专业的文档\u003C\u002Fstrong>\n\u003C\u002Fp>\n\n\u003Chr \u002F>\n\n# 😺 为什么选择Litho\n\n- **自动保持文档与代码库同步**，不再有陈旧过时的文档\n- **节省数百小时**的手动文档编写和维护时间\n- **改善新成员入职体验**，提供全面、最新的文档支持\n- **提升代码评审效率**，通过清晰的架构背景信息辅助决策\n- **满足合规性要求**，提供可审计的自动化文档\n- **支持多种编程语言**（Rust、Python、Java、Go、C#、JavaScript等）\n- **生成专业的C4模型图**，涵盖上下文、容器、组件及代码细节\n- **集成CI\u002FCD流水线**，实现每次提交自动更新文档\n\n🌟 **适用人群：**\n- 各规模的开发团队\n- 开源项目\n- 企业级软件开发人员\n- 所有讨厌维护过时文档的人！\n\n❤️ 如果您喜欢**Litho**，请给它点个星🌟，或者[赞助我](https:\u002F\u002Fgithub.com\u002Fsponsors\u002Fsopaco)！❤️\n\n**感谢各位的支持**\n\n[![Stargazers repo roster for @sopaco\u002Fdeepwiki-rs](https:\u002F\u002Freporoster.com\u002Fstars\u002Fsopaco\u002Fdeepwiki-rs)](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fstargazers)\n\n# 🌠 功能与特性\n\n### 核心能力\n- 基于代码分析的AI驱动架构文档生成\n- 自动创建C4模型图（上下文、容器、组件、代码）\n- 智能提取代码注释、结构和关系\n- 多语言支持，覆盖多种编程语言\n- 可定制的文档输出模板系统\n\n### 高级特性\n- **外部知识整合**：挂载外部文档（PDF、Markdown、SQL等）作为知识源，增强分析能力\n- **数据库文档**：为SQL项目自动生成包含ERD图的数据库模式文档\n- Git历史分析，追踪架构演进过程\n- 代码元素与文档之间的交叉引用\n- 嵌入式图表和示例的交互式文档\n- 与CI\u002FCD流水线集成，实现自动化文档生成\n\n## 💡 解决的问题\nLitho解决了技术文档过时、不完整这一常见问题，通过自动从源代码中生成最新架构文档来实现。再也不用担心手动维护的文档会落后于代码变更——Litho让您的文档始终与代码库保持一致。\n\n# 🌐 Litho生态体系\nLitho是更广泛工具生态系统的一部分，旨在提升开发者的生产力和文档质量。Litho生态体系包含一系列与Litho无缝协作的互补工具，共同构成完整的文档工作流：\n\n## 📘 Litho Book\n**Litho Book**是一款基于Rust和Axum构建的高性能Markdown阅读器，专为优雅地浏览Litho生成的文档而设计。\n\n### 核心功能\n- 实时 Markdown 渲染，支持语法高亮\n- 完整支持 Mermaid 图表，用于架构图绘制\n- 智能搜索，支持文件和内容的模糊匹配\n- 高性能架构，内存占用低\n- AI 智能文档解析与问答\n\n### 🌠 截图\n\u003Cdiv style=\"text-align: center;\">\n  \u003Ctable style=\"width: 100%; margin: 0 auto;\">\n    \u003Ctr>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_ea6525878d02.webp\" alt=\"snapshot-1\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_5ec994c985be.webp\" alt=\"snapshot-2\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n    \u003C\u002Ftr>\n  \u003C\u002Ftable>\n\u003C\u002Fdiv>\n\n### 与 Litho 的集成\nLitho Book 是消费 Litho 生成文档的理想配套应用。典型的工作流程如下：\n1. 使用 Litho 从代码库中生成文档\n2. 使用 Litho Book 通过优雅的界面浏览和探索生成的文档\n\n[了解更多关于 Litho Book 的信息](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Flitho-book)\n\n## 🔧 Mermaid Fixer\n**Mermaid Fixer** 是一款高性能的 AI 驱动工具，能够自动检测并修复 Markdown 文件中 Mermaid 图表的语法错误。\n\n### 核心功能\n- 自动扫描目录中的 Markdown 文件\n- 使用 JS 沙箱验证精准检测 Mermaid 语法错误\n- 基于 LLM 的智能修复\n- 提供修复前后的全面报告\n- 灵活的配置，支持多种 LLM 提供商\n\n### 与 Litho 的集成\nMermaid Fixer 通过自动修复 Mermaid 图表中的语法错误，提升了 Litho 生成文档的质量。这确保了文档中的所有架构图都是有效的，并能正确渲染。\n\n### 👀 截图\n\u003Cdiv style=\"text-align: center;\">\n  \u003Ctable style=\"width: 100%; margin: 0 auto;\">\n    \u003Ctr>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_ec6d4ce10b4d.webp\" alt=\"snapshot-1\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n      \u003Ctd style=\"width: 50%;\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_9b9f66493fda.webp\" alt=\"snapshot-2\" style=\"width: 100%; height: auto; display: block;\">\u003C\u002Ftd>\n    \u003C\u002Ftr>\n  \u003C\u002Ftable>\n\u003C\u002Fdiv>\n\n[了解更多关于 Mermaid Fixer 的信息](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fmermaid-fixer)\n\n## 🤖Agent 技能\n在 Smithery 中运行！ [![在 Smithery 中运行](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_readme_cded48b327c8.png)](https:\u002F\u002Fsmithery.ai\u002Fskills?ns=sopaco&utm_source=github&utm_medium=badge)\n\n# 🧠 工作原理\n[![zread](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAsk_Zread-_.svg?style=flat&color=00b0aa&labelColor=000000&logo=data%3Aimage%2Fsvg%2Bxml%3Bbase64%2CPHN2ZyB3aWR0aD0iMTYiIGhlaWdodD0iMTYiIHZpZXdCb3g9IjAgMCAxNiAxNiIgZmlsbD0ibm9uZSIgeG1sbnM9Imh0dHA6Ly93d3cudzMub3JnLzIwMDAvc3ZnIj4KPHBhdGggZD0iTTQuOTYxNTYgMS42MDAxSDIuMjQxNTZDMS44ODgxIDEuNjAwMSAxLjYwMTU2IDEuODg2NjQgMS42MDE1NiAyLjI0MDFVNC45NjAxQzEuNjAxNTYgNS4zMTM1NiAxLjg4ODEgNS42MDAxIDIuMjQxNTYgNS42MDAxSDQuOTYxNTZDNS4zMTUwMiA1LjYwMDEgNS42MDE1NiA1LjMxMzU2IDUuNjAxNTYgNC45NjAxVjIuMjQwMUM1LjYwMTU2IDEuODg2NjQgNS4zMTUwMiAxLjYwMDEgNC45NjE1NiAxLjYwMDFaIiBmaWxsPSIjZmZmIi8%2BCjxwYXRoIGQ9Ik00Ljk6MTU2IDEwLjM5OTlIMi4yNDE1NkC1Ljg4ODEgMTAuMzk5OSAxLjYwMTU2IDEwLjY4NjQgMS42MDE1NiAxMS1wMzk5VjEzLjc1OTlDMS42MDE1NiAxNC4xMTM0IDEuODg4MSAxNC4zOTk5IDIuMjQxNTYgMTQuMzk5OUH0Ljk6MTU2QzUuMzE1MDIgMTQuMzk9OSA1LjYwMTU2AxNC4xMTM0IDEuODg6NjQgNS4zE1MDIgMTAuMzk9OSA0Ljk6MTU2IDEwLjM9OTLaIiBmaWxsPSIjZmZmIi8%2BCjxwYXRoIGQ9Ik0xMy7584IDEuNjAwMUgxLSwMzg4QzEwLjY8NSIDEuNjAwDEgMTAuMzk8CAxLjg8NCAxLjM9OD4gMTAuMzk8CAxLSwMzg4CAxMy7584QzE4LjExMTkgNS42MDAxIDE4LjM9OD4gNS4zE1MDIgIDE4LjM9OD4gNC49T60V2LjI0MDFDMT49OD4gIDEuODg6NjQgIDE4LjExMTkgIDEuNjAwDEgMy7584IDEuNjAwDEWiIgZmlsbD0iI2ZmZiIvPgo8cGF0aCBkPSJNNCAxMkwxMiA0TDQgMTJaIiBmaWxsPSIjZmZmIi8%2BCjxwYXRoIGQ9Ik00IDEyTDEyIDQiIHN0cm9rZT0iI2ZmZiIgc3Ryb2tlLXdpZHRoPSIxLjUiIHN0cm9rZS1saW5lY2FwPSJyb3VuZCIvPgo8L3N2Zz4K&logoColor=ffffff)](https:\u002F\u002Fzread.ai\u002Fsopaco\u002Fdeepwiki-rs)\n\n## 四阶段处理流程\nLitho 的架构围绕一个四阶段处理流程设计，该流程将原始代码转换为全面的文档：\n\n```mermaid\nflowchart TD\n    A[输入：源代码仓库] --> B[阶段 1：预处理]\n    B --> C[阶段 2：智能研究与分析]\n    C --> D[阶段 3：文档生成]\n    D --> E[阶段 4：验证与增强]\n    E --> F[输出：高质量技术文档]\n\n    subgraph 预处理阶段\n        B1[代码扫描与发现]\n        B2[多语言语法分析]\n        B3[结构与依赖提取]\n        B4[代码洞察生成]\n        B5[Agent 内存块初始化]\n        B --> B1 --> B2 --> B3 --> B4 --> B5\n    end\n\n    subgraph 智能研究与分析阶段\n        C1[系统上下文研究员]\n        C2[领域模块检测器]\n        C3[工作流研究员]\n        C4[边界分析仪]\n        C5[关键模块洞察官]\n        C6[Agent 内存块读写]\n        C7[ReAct 推理循环]\n        C --> C1 --> C2 --> C3 --> C4 --> C5 --> C6 --> C7\n    end\n\n    subgraph 文档生成阶段\n        D1[概览文档编辑器]\n        D2[架构文档编辑器]\n        D3[工作流文档编辑器]\n        D4[边界文档编辑器]\n        D5[关键模块编辑器]\n        D6[Agent 内存块读取]\n        D7[高质量文档组装]\n        D --> D1 --> D2 --> D3 --> D4 --> D5 --> D6 --> D7\n    end\n\n    subgraph 验证与增强阶段\n        E1[Mermaid 语法验证]\n        E2[文档完整性检查]\n        E3[图表自动修复]\n        E4[质量报告生成]\n        E5[最终文档输出]\n        E --> E1 --> E2 --> E3 --> E4 --> E5\n    end\n\n    style B fill:#e3f2fd,stroke:#1976d2\n    style C fill:#f3e5f5,stroke:#7b1fa2\n    style D fill:#e8f5e8,stroke:#388e3c\n    style E fill:#fff3e0,stroke:#e65100\n```\n\n### 预处理阶段\nLitho 首先会扫描整个代码库，以识别源文件、提取元数据并分析项目结构。此阶段：\n- 发现多种语言的所有源代码文件\n- 解析文件结构并识别关键组件\n- 提取注释、文档字符串和代码标注\n- 识别模块和组件之间的依赖关系\n- 构建代码库的全面表示\n\n```mermaid\nflowchart TD\nA[预处理代理] --> B[结构提取器]\nA --> C[原始文档提取器]\nA --> D[代码分析代理]\nA --> E[关系分析代理]\nB --> F[项目结构]\nC --> G[原始文档材料]\nD --> H[核心代码洞察]\nE --> I[代码依赖关系]\nF --> J[存储到内存]\nG --> J\nH --> J\nI --> J\n```\n\n### 研究阶段\n在这个由 AI 驱动的阶段，Litho 会分析代码结构以理解架构意图：\n- 应用机器学习模型来识别模式和关系\n- 根据代码结构和命名规范推断架构角色\n- 确定组件边界和服务职责\n- 绘制组件之间的依赖关系和数据流\n- 识别潜在的架构异味和反模式\n- 为每个组件生成上下文感知的文档\n\n```mermaid\nflowchart TD\nA[研究编排器] --> B[系统上下文研究员]\nA --> C[领域模块检测器]\nA --> D[架构研究员]\nA --> E[工作流研究员]\nA --> F[关键模块洞察]\nB --> G[系统上下文报告]\nC --> H[领域模块报告]\nD --> I[架构分析报告]\nE --> J[工作流分析报告]\nF --> K[模块深度洞察]\nG --> 内存\nH --> 内存\nI --> 内存\nJ --> 内存\nK --> 内存\n```\n\n### 组合与输出阶段\nLitho 将分析后的信息组合成结构化的文档格式：\n- 生成 C4 模型图（上下文、容器、组件、代码）\n- 创建具有清晰导航的层次化文档结构\n- 嵌入相关的代码示例和解释\n- 对所有文档应用一致的样式和格式\n- 添加相关组件和图表之间的交叉引用\n\n```mermaid\nflowchart TD\nA[文档编排器] --> B[概览编辑器]\nA --> C[架构编辑器]\nA --> D[模块洞察编辑器]\nB --> E[概述文档]\nC --> F[架构文档]\nD --> G[模块文档]\nE --> H[文档树]\nF --> H\nG --> H\nH --> I[磁盘输出]\nI --> J[输出目录]\n```\n\n### 验证与增强阶段\n最后阶段确保文档的质量和完整性：\n- 验证图表语法和一致性\n- 检查文档覆盖是否完整\n- 识别文档中的空白并提出改进建议\n- 与 Mermaid Fixer 集成，确保所有图表正确渲染\n- 生成文档覆盖率的统计和报告\n- 创建索引和目录以便于导航\n\n# 🏗️ 架构概述\n\n**Litho** 具有复杂的模块化架构设计，旨在实现高性能、可扩展性和智能分析。该系统采用多阶段工作流程，配备专门的 AI 代理和全面的缓存机制。\n\n```mermaid\ngraph LR\n    subgraph 输入阶段\n        A[CLI 启动] --> B[加载配置]\n        B --> C[扫描结构]\n        C --> D[提取 README]\n    end\n    subgraph 分析阶段\n        D --> E[语言解析]\n        E --> F[AI 增强分析]\n        F --> G[存储到内存]\n    end\n    subgraph 推理阶段\n        G --> H[编排器启动]\n        H --> I[系统上下文分析]\n        H --> J[领域模块检测]\n        H --> K[工作流分析]\n        H --> L[关键模块洞察]\n        I --> M[存储到内存]\n        J --> M\n        K --> M\n        L --> M\n    end\n    subgraph 编排阶段\n        M --> N[编排中心启动]\n        N --> O[生成项目概览]\n        N --> P[生成架构图]\n        N --> Q[生成工作流文档]\n        N --> R[生成模块洞察]\n        O --> S[写入 DocTree]\n        P --> S\n        Q --> S\n        R --> S\n    end\n    subgraph 输出阶段\n        S --> T[持久化文档]\n        T --> U[生成总结报告]\n    end\n```\n\n## 核心模块\nLitho 的架构由多个相互连接的模块组成，协同工作以实现无缝的文档生成：\n\n- **代码扫描器**：发现并分析多种语言的源代码文件\n- **语言解析器**：使用特定于语言的解析器从代码中提取结构信息\n- **架构分析器**：基于 AI 的组件，用于推断架构模式和关系\n- **图表生成器**：使用 Mermaid 语法创建 C4 模型图\n- **文档格式化器**：将内容组织成结构化的可导航文档\n\n## 核心流程\n核心处理流程遵循确定性的管道：\n1. **扫描** - 发现并分析源代码文件\n2. **解析** - 提取结构和语义信息\n3. **分析** - 应用 AI 模型推断架构和关系\n4. **生成** - 创建图表和文档内容\n5. **格式化** - 将内容组织成结构化的文档\n6. **导出** - 以所需格式输出\n\n```mermaid\nsequenceDiagram\nparticipant Main as main.rs\nparticipant Workflow as workflow.rs\nparticipant Context as GeneratorContext\nparticipant Preprocess as PreProcessAgent\nparticipant Research as ResearchOrchestrator\nparticipant Doc as DocumentationOrchestrator\nparticipant Outlet as DiskOutlet\nMain->>Workflow : launch(config)\nWorkflow->>Context : Create context (LLM, Cache, Memory)\nWorkflow->>Preprocess : execute(context)\nPreprocess->>Context : Store project structure and metadata\nContext-->>Workflow : Preprocessing complete\nWorkflow->>Research : execute_research_pipeline(context)\nResearch->>Research : Execute multiple research agents in parallel\nloop Each Research Agent\nResearch->>StepForwardAgent : execute(context)\nStepForwardAgent->>Context : Validate data sources\nStepForwardAgent->>AgentExecutor : Call prompt or extract\nAgentExecutor->>LLMClient : Initiate LLM request\nLLMClient->>CacheManager : Check cache\nalt Cache hit\nCacheManager-->>LLMClient : Return cached result\nelse Cache miss\nLLMClient->>LLM : Call LLM API\nLLM-->>LLMClient : Return raw response\nLLMClient->>CacheManager : Store result to cache\nend\nLLMClient-->>AgentExecutor : Return processed result\nAgentExecutor-->>StepForwardAgent : Return result\nStepForwardAgent->>Context : Store result to Memory\nend\nResearch-->>Workflow : Research complete\nWorkflow->>Doc : execute(context, doc_tree)\nDoc->>Doc : Call multiple composition agents to generate docs\nDoc-->>Workflow : Documentation generation complete\nWorkflow->>Outlet : save(context)\nOutlet-->>Workflow : Storage complete\nWorkflow-->>Main : Process finished\n```\n\n# 🖥 开始使用\n\n### 前置条件\n- [**Rust**](https:\u002F\u002Fwww.rust-lang.org)（版本 1.70 或更高）\n- [**Cargo**](https:\u002F\u002Fdoc.rust-lang.org\u002Fcargo\u002F)\n\n### 安装\n#### 选项 1：从 crates.io 安装（推荐）\n```sh\ncargo install deepwiki-rs\n```\n\n#### 选项 2：从源码构建\n1. 克隆仓库：\n    ```sh\n    git clone https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs.git\n    ```\n2. 进入项目目录：\n    ```sh\n    cd deepwiki-rs\n    ```\n3. 构建项目：\n    ```sh\n    cargo build --release\n    ```\n4. 编译后的二进制文件将位于 `target\u002Frelease` 目录中。\n\n# 🚀 使用方法\n**Litho** 提供了一个简单的命令行界面，用于从您的代码库生成文档。更多配置参数，请参阅 [CLI 选项详情](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fblob\u002Fmain\u002Fdocs\u002F5%E3%80%81%E8%BE%B9%E7%95%8C%E8%B0%83%E7%94%A8.md#litho)。\n\n### 基本命令\n```sh\ndeepwiki-rs -p .\u002Fmy-project -o .\u002Fdocs\n\n# 以目标语言生成文档。\ndeepwiki-rs --target-language en -p .\u002Fmy-project\n\ndeepwiki-rs --target-language ja -p .\u002Fmy-project\n```\n\n此命令将：\n- 扫描 `.\u002Fmy-project` 中的所有文件\n- 分析代码结构和关系\n- 生成全面的 C4 架构文档\n- 将输出保存到 `.\u002Flitho.docs` 目录。\n\n### 文档生成\nLitho 支持多种文档生成选项：\n\n```sh\n# 使用默认设置生成文档\ndeepwiki-rs 跳过生成流程中的某些处理阶段\ndeepwiki-rs --skip-preprocessing --skip-research\n```\n\n### 高级选项\n```sh\n# 关闭 ReAct 模式，避免通过工具调用自动扫描项目文件\ndeepwiki-rs -p .\u002Fsrc --disable-preset-tools --llm-api-base-url \u003Cyour llm provider base-api> --llm-api-key \u003Cyour api key> --model-efficient GPT-5-mini\n\n# 同时设置高效模型和强大模型\ndeepwiki-rs -p .\u002Fsrc --model-efficient GPT-5-mini --model-poweruful GPT-5-Pro --llm-api-base-url \u003Cyour llm provider base-api> --llm_api_key \u003Cyour api key> --model-efficient GPT-5-mini\n```\n\n## 📚 外部知识集成\n\nLitho 支持挂载外部文档作为知识源，以增强生成文档的业务背景和架构决策信息。\n\n### 支持的文档类型\n- **PDF** - 架构图、设计文档\n- **Markdown** - 技术文档、ADR\n- **SQL** - 数据库模式文件\n- **YAML\u002FJSON** - API 规范（OpenAPI）、配置文件\n- **文本** - 纯文本文档\n\n### 知识类别\n文档被组织成不同的类别，以便有针对性地传递给特定的代理：\n- `architecture` - 系统架构和 C4 模型文档\n- `database` - 模式、ERD 和数据模型文档\n- `api` - API 规范和端点文档\n- `deployment` - 基础设施和 DevOps 文档\n- `adr` - 架构决策记录\n- `workflow` - 业务流程和工作流\n- `general` - 未分类的一般文档\n\n### 同步知识命令\n```sh\n# 同步外部知识源（处理并缓存本地文档）\ndeepwiki-rs sync-knowledge\n\n# 即使缓存最新也强制同步\ndeepwiki-rs sync-knowledge --force\n```\n\n### 配置示例（litho.toml）\n```toml\n[knowledge.local_docs]\nenabled = true\ncache_dir = \".litho\u002Fcache\u002Fknowledge\u002Flocal_docs\"\nwatch_for_changes = true\n\n# 大型文档的默认分块设置\n[knowledge.local_docs.default_chunking]\nenabled = true\nmax_chunk_size = 8000\nchunk_overlap = 200\nstrategy = \"semantic\"  # 选项：语义、段落、固定\nmin_size_for_chunking = 10000\n\n# 架构文档类别\n[[knowledge.local_docs.categories]]\nname = \"architecture\"\ndescription = \"系统架构文档\"\npaths = [\n    \"docs\u002Farchitecture\u002F**\u002F*.md\",\n    \"docs\u002Fdesign\u002F**\u002F*.pdf\"\n]\ntarget_agents = [\n    \"SystemContextResearcher\",\n    \"ArchitectureResearcher\",\n    \"ArchitectureEditor\"\n]\n\n# 数据库文档类别\n[[knowledge.local_docs.categories]]\nname = \"database\"\ndescription = \"数据库模式文档\"\npaths = [\n    \"docs\u002Fdatabase\u002F**\u002F*.md\",\n    \"docs\u002Fschema\u002F**\u002F*.sql\"\n]\ntarget_agents = [\n    \"ArchitectureResearcher\",\n    \"DomainModulesDetector\",\n    \"KeyModulesInsight\"\n]\n```\n\n## 🗄️ 数据库文档\n\nLitho 会自动分析 SQL 数据库项目（`.sqlproj`）和 SQL 文件，生成全面的数据库文档，包括：\n\n- **数据库项目** - SQL Server 项目结构\n- **表** - 模式、列、数据类型、约束、主键\n- **视图** - 视图定义及引用的表\n- **存储过程** - 参数、操作、访问的表\n- **函数** - 标量函数和表值函数\n- **关系** - 外键和隐式引用（附带 ERD 图）\n- **数据流** - ETL 操作和数据流动模式\n\n### 数据库分析功能\n```\n📊 数据库代码分布：项目(2) SQL 文件(15) DAO(3)\n✅ 数据库概览分析已完成：\n   - 数据库项目：2 项\n   - 表：12 项\n   - 视图：5 项\n   - 存储过程：8 项\n   - 函数：3 项\n   - 表关系：6 项\n   - 数据流：4 项\n   - 置信度：8.5\u002F10\n```\n\n### 生成的数据库文档\n数据库文档会自动包含在输出中，名为 `6.Database-Overview.md`，内容包括：\n- 概要统计表\n- 包含列定义的详细表模式\n- 展示关系的 Mermaid ER 图\n- 存储过程文档\n- 数据流动描述\n\n## 📁 输出结构\nLitho 会生成一个组织良好的文档结构：\n\n```\nproject-docs\u002F\n├── 1. Project Overview      # 项目概述、核心功能、技术栈\n├── 2. Architecture Overview # 整体架构、核心模块、模块分解\n├── 3. Workflow Overview     # 整体流程、核心流程\n├── 4. Deep Dive\u002F            # 详细的技术主题实现文档\n│   ├── Topic1.md\n│   ├── Topic2.md\n├── 5. Boundary-Interfaces   # API 端点、外部集成\n├── 6. Database-Overview     # 数据库模式、表、关系（仅限 SQL 项目）\n```\n\n# 🤝 贡献\n我们欢迎任何形式的贡献！请通过 [GitHub Issues](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fissues) 报告错误或提交功能请求。\n\n## 贡献方式\n- **语言支持**：增加对其他编程语言的支持\n- **模板创建**：设计新的文档模板和样式\n- **图表优化**：改进 Mermaid 图表生成算法\n- **性能优化**：提升处理速度和内存使用效率\n- **测试覆盖**：为各种代码模式添加全面的测试用例\n- **文档编写**：完善项目文档和使用指南\n- **Bug 修复**：帮助识别并修复代码库中的问题\n\n## 开发贡献流程\n1. Fork 该项目\n2. 创建一个特性分支（`git checkout -b feature\u002Famazing-feature`）\n3. 提交你的更改（`git commit -m '添加一项超赞功能'`）\n4. 推送到该分支（`git push origin feature\u002Famazing-feature`）\n5. 创建一个 Pull Request\n\n# 🪪 许可证\n**MIT**。许可证副本已在 [LICENSE](LICENSE) 文件中提供。\n\n# 👨 关于我\n> 🚀 通过在 [GitHub 上赞助](https:\u002F\u002Fgithub.com\u002Fsponsors\u002Fsopaco) 来帮助我更好地开发这款软件\n\n作为一名经验丰富的互联网老兵，我经历了个人电脑互联网、移动互联网和人工智能应用的浪潮。从一名独立的移动应用开发者到企业界的从业者，我在产品设计与研发方面积累了丰富的经验。目前就职于 [快手](https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FKuaishou)，专注于通用前端系统的研发及人工智能领域的探索。\n\nGitHub：[sopaco](https:\u002F\u002Fgithub.com\u002Fsopaco)","# Litho (deepwiki-rs) 快速上手指南\n\nLitho 是一个基于 Rust 构建的高性能 AI 驱动文档生成引擎。它能自动分析源代码，生成符合 C4 模型的专业架构文档（包含上下文图、容器图、组件图等），让文档与代码始终保持同步。\n\n## 环境准备\n\n在开始之前，请确保您的开发环境满足以下要求：\n\n*   **操作系统**: Linux, macOS, 或 Windows (WSL 推荐)\n*   **Rust 工具链**: 需安装 Rust (建议最新稳定版)\n    *   安装命令：`curl --proto '=https' --tlsv1.2 -sSf https:\u002F\u002Fsh.rustup.rs | sh`\n    *   国内加速安装：`export RUSTUP_DIST_SERVER=https:\u002F\u002Frsproxy.cn && export RUSTUP_UPDATE_ROOT=https:\u002F\u002Frsproxy.cn\u002Frustup && curl --proto '=https' --tlsv1.2 -sSf https:\u002F\u002Frsproxy.cn\u002Frustup-init.sh | sh`\n*   **AI 模型访问**: 需要配置可用的 LLM API Key (如 OpenAI, Azure, 或兼容接口)，用于智能分析与绘图。\n*   **目标代码库**: 准备好您想要生成文档的源代码仓库。\n\n## 安装步骤\n\n您可以通过 Cargo 直接安装 Litho 命令行工具。\n\n### 使用官方源安装\n```bash\ncargo install deepwiki-rs\n```\n\n### 使用国内镜像源安装 (推荐)\n如果您在中国大陆，建议使用清华或中科大镜像源以加速下载：\n\n```bash\n# 临时设置镜像源并安装\nexport CARGO_REGISTRIES_CRATES_IO_PROTOCOL=sparse\nexport CARGO_REGISTRIES_CRATES_IO_INDEX=https:\u002F\u002Fmirrors.tuna.tsinghua.edu.cn\u002Fgit\u002Fcrates.io-index.git\ncargo install deepwiki-rs\n```\n\n安装完成后，验证版本：\n```bash\ndeepwiki-rs --version\n```\n\n## 基本使用\n\nLitho 的核心功能是将任意代码库转换为结构化的 Markdown 文档和 Mermaid 图表。\n\n### 1. 基础生成命令\n\n进入您的项目根目录，运行以下命令即可开始生成文档：\n\n```bash\ndeepwiki-rs generate .\n```\n\n*   `.` 表示当前目录作为源代码输入。\n*   默认情况下，生成的文档将输出到 `.\u002Flitho-output` 目录（具体路径视版本配置而定）。\n\n### 2. 指定输出目录与模型配置\n\n您可以自定义输出路径并传入必要的 AI 配置：\n\n```bash\ndeepwiki-rs generate .\u002Fmy-project \\\n  --output .\u002Fdocs-generated \\\n  --model gpt-4o \\\n  --api-key $YOUR_OPENAI_API_KEY\n```\n\n**参数说明：**\n*   `generate \u003Cpath>`: 指定要分析的源代码路径。\n*   `--output \u003Cdir>`: 指定文档生成的输出目录。\n*   `--model \u003Cname>`: 指定使用的 LLM 模型名称。\n*   `--api-key \u003Ckey>`: 您的 API 密钥（也可通过环境变量 `OPENAI_API_KEY` 设置）。\n\n### 3. 查看生成的文档\n\n生成完成后，您将得到一套完整的 Markdown 文件，其中包含自动绘制的 Mermaid 架构图。\n\n**推荐搭配 Litho Book 查看：**\n为了获得最佳阅读体验（支持实时渲染 Mermaid 图表、智能搜索），建议使用配套工具 **Litho Book** 浏览生成的文档：\n\n```bash\n# 假设已安装 litho-book\nlitho-book .\u002Fdocs-generated\n```\n\n如果没有安装 Litho Book，您也可以直接使用任何支持 Mermaid 的 Markdown 编辑器（如 VS Code + Mermaid 插件）打开生成的 `.md` 文件进行预览。\n\n---\n**提示**: 首次运行可能需要几分钟时间，具体取决于代码库大小和所选模型的响应速度。生成的文档包含 C4 模型的四个层级（Context, Container, Component, Code），可直接用于团队知识库或集成到 CI\u002FCD 流程中。","某金融科技公司后端团队在接手一个遗留的微服务支付系统时，急需理清复杂的模块依赖关系以进行重构，但原有文档已严重滞后于代码现状。\n\n### 没有 deepwiki-rs 时\n- 架构文档停留在两年前的版本，与新上线的支付网关和风控模块完全脱节，开发人员只能靠“猜”来理解代码逻辑。\n- 手动绘制 C4 架构图耗时费力，每次代码迭代后图表迅速过时，导致技术评审会上因信息不准而频繁争论。\n- 新入职工程师需要花费数周时间阅读源码才能梳理出数据流向，严重拖慢了故障排查和功能开发进度。\n- 缺乏统一的文档结构，不同成员编写的说明文件格式混乱，难以形成系统性的知识库供 AI 助手学习。\n\n### 使用 deepwiki-rs 后\n- deepwiki-rs 自动扫描最新代码库，瞬间生成与当前生产环境完全同步的精准架构文档，消除了信息偏差。\n- 基于 C4 模型自动生成上下文图、容器图和组件图，将原本需要数天的绘图工作缩短至几分钟，且随代码提交实时更新。\n- 清晰的层级结构和自动提取的代码级文档，让新成员能在一天内掌握核心支付链路，大幅降低上手门槛。\n- 输出标准化、结构化的技术内容，不仅便于人类团队查阅，更为内部 AI 编程助手提供了高质量的上下文语境。\n\ndeepwiki-rs 将原本滞后的文档负担转化为实时的架构资产，让团队能专注于业务创新而非维护过时的说明书。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsopaco_deepwiki-rs_5ec994c9.webp","sopaco","Sopaco","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fsopaco_facc619c.jpg","@快手 姜萌","BeiJing","dokhell@live.cn",null,"https:\u002F\u002Fmemclaw.netlify.app\u002F","https:\u002F\u002Fgithub.com\u002Fsopaco",[24,28,32],{"name":25,"color":26,"percentage":27},"Rust","#dea584",97.4,{"name":29,"color":30,"percentage":31},"Shell","#89e051",2.2,{"name":33,"color":34,"percentage":35},"Go Template","#00ADD8",0.4,868,109,"2026-04-07T01:48:02","MIT",2,"未说明",{"notes":43,"python":44,"dependencies":45},"该工具名为 deepwiki-rs (Litho)，是使用 Rust 编写的高性能 AI 驱动文档生成引擎，并非 Python 项目。它本身不包含模型，需要用户自行配置并连接外部大语言模型 (LLM) 提供商。配套有 Litho Book (文档浏览器) 和 Mermaid Fixer (图表修复工具)。支持多种编程语言代码库的分析。","不适用 (基于 Rust)",[46,47,48],"Rust (语言基础)","Axum (用于 Litho Book)","LLM Provider (需配置外部大模型服务)",[50,51],"开发框架","语言模型",[53,54,55,56,57,58,59,60,61],"deepwiki","llm","rust","claude","deepseek","mistral","openai","openrouter","wiki","ready","2026-03-27T02:49:30.150509","2026-04-07T14:37:09.371060",[66,71,76,81,86,90],{"id":67,"question_zh":68,"answer_zh":69,"source_url":70},22273,"为什么 AI 分析阶段一直失败且提示 'No data extracted'？","这通常是由于模型兼容性不足或本地代理配置问题导致的。\n1. 确保已升级到最新版本（v1.5.0+）。\n2. 检查本地 API 代理（如通过 aichat 代理 GitHub Copilot）是否正确处理了流式响应或 JSON 格式。\n3. 尝试直接在命令行指定更强大的模型组合，例如：`--model-powerful github-copilot:claude-sonnet-4.6`。","https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fissues\u002F89",{"id":72,"question_zh":73,"answer_zh":74,"source_url":75},22268,"遇到 'Failed to deserialize the extracted data: invalid type: string' 或无限循环报错怎么办？","这是模型兼容性问题，通常发生在旧版本中。请升级到 deepwiki-rs v1.5.0 或最新代码，该版本显著增强了模型兼容性并解决了此类反序列化错误。\n更新命令参考：\ngit pull origin main\n# 或者下载最新 release: https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Freleases\u002Ftag\u002F1.5.0","https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fissues\u002F83",{"id":77,"question_zh":78,"answer_zh":79,"source_url":80},22269,"推荐使用哪些大语言模型以获得最佳效果？","根据维护者建议，不同地区的推荐模型如下：\n1. 国内用户：优先尝试 Kimi 2.5（指令遵循能力极佳），也可使用 Qwen3-Next-80B-A3B-Instruct 或 DeepSeek 3.1 及更新版本。\n2. 国外用户：推荐使用 OpenAI、Claude 或 Gemini 系列模型。\n注意：部分小型化自部署模型可能会导致持续的解析错误。","https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fissues\u002F47",{"id":82,"question_zh":83,"answer_zh":84,"source_url":85},22270,"调用模型时出现 'unknown variant' 错误或反序列化失败如何解决？","这通常是因为模型输出的格式不符合预期枚举值（如 code_purpose 字段）。\n解决方案：\n1. 升级软件至 v1.5.0 及以上版本，该版本提升了容错率和模型兼容性。\n2. 尝试更换模型，用户反馈 DeepSeek-v3.2 表现较好，而某些版本的 GPT-5-mini 可能存在问题。\n3. 如果是自部署的小型模型，建议切换回主流大模型 API。","https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fissues\u002F20",{"id":87,"question_zh":88,"answer_zh":89,"source_url":70},22271,"遇到 'data did not match any variant of untagged enum ApiResponse' 错误怎么处理？","此错误表示 API 返回的数据格式无法被程序识别。请执行以下操作：\n1. 立即升级到 v1.5.0 版本或拉取最新主分支代码，相关修复已合并（PR #96, #94）。\n2. 检查使用的本地代理（如 aichat --serve）是否正确转发了响应，确保没有额外的包装层导致 JSON 结构变化。",{"id":91,"question_zh":92,"answer_zh":93,"source_url":80},22272,"使用特定参数（如 --llm-provider openai）时报错该如何调整？","如果遇到反序列化错误，尝试移除显式的 provider 参数。有用户反馈删除 `--llm-provider openai` 参数后问题解决，让程序自动检测或使用默认配置可能更稳定。",[95,100,105,110,115,120,125,130,135,140,145,150,155],{"id":96,"version":97,"summary_zh":98,"released_at":99},135996,"1.5.0","### 🔥 重大修复：LLM 反序列化可靠性\n\n**本次发布解决了长期困扰不同 LLM 提供商生产环境使用的问题。**\n\n#### 问题描述\n此前，Litho 依赖 LLM 严格遵循 JSON Schema 并返回格式完美的结构化 JSON。然而，在实际项目中，面对复杂的任务和多种模型（尤其是非 OpenAI 提供商如 Ollama、DeepSeek 等），这种做法经常导致解析错误和处理失败。系统过于脆弱，一旦模型输出偏离预期格式，就会彻底中断流程。\n\n#### 解决方案\n我们实施了一套全面的多层机制，以确保结构化数据提取的可靠性：\n\n**1. 宽松反序列化与智能回退**  \n- 增加了健壮的回退处理逻辑，可优雅地应对格式错误的 JSON 响应  \n- 支持多种解析策略：严格模式 → 松散模式 → 文本提取  \n- 在保持系统运行的同时，提供详细的错误上下文以便调试  \n- 自动修复常见的 LLM 格式化错误  \n\n**2. 针对不同提供商的专用提取器**  \n- **OllamaExtractorWrapper**：针对 Ollama 及类似仅返回文本的模型，新增从 Markdown 代码块中智能提取 JSON 的功能，并加入重试机制  \n- **OpenAICompatibleExtractorWrapper**：增强对 OpenAI 兼容 API 的支持，在可用时优先使用结构化输出  \n- 统一接口可根据提供商自动选择合适的提取策略  \n\n**3. 提升提示工程设计**  \n- 优化系统提示，更强调 JSON 格式要求  \n- 针对边缘场景（空值、可选字段、数组格式等）添加具体指令  \n- 更清晰地指导嵌套对象结构的构建  \n\n**4. 可配置的重试与退避机制**  \n- 针对速率限制采用带有抖动的指数退避策略  \n- 每个提供商的尝试次数可通过 `llm.retry_attempts` 进行配置  \n- 详细记录失败和重试情况  \n\n**影响**：用户在以下场景中将显著减少失败次数：  \n- 使用本地模型（Ollama、Llama 等）  \n- 使用其他云服务提供商（DeepSeek、Moonshot、Mistral、OpenRouter 等）  \n- 处理包含嵌套结构的复杂 Schema  \n- 分析大型代码库  \n\n### 🗄️ 数据库文档生成（新功能）\n\n- 引入 `DatabaseOverviewAnalyzer` 代理  \n- 能够分析 SQL 数据库项目中的表、视图、存储过程及关系  \n- 自动生成包含表结构关系的完整数据库文档  \n- 支持按知识类别筛选，实现上下文感知的分析  \n- 特别适用于 SQL Server 和数据仓库项目  \n\n### 📄 外部知识集成（新功能）\n\n新增强大的 `local_docs` 集成功能，用于导入外部文档：\n\n**支持的文件类型**：PDF、Markdown、Text、SQL、YAML、JSON  \n\n**核心能力**：  \n- 按类别组织文档（例如：“架构”、“API”、“数据库”）  \n- 实现定向分发：可将文档分配给特定代理  \n- 可配置的分块策略：","2026-04-05T06:43:39",{"id":101,"version":102,"summary_zh":103,"released_at":104},135997,"1.3.0","🚀 v1.3.0 新功能\n> 感谢贡献者 @MrLYC、@vovanduc 和 @gurdasnijor\n> **完整变更日志**：https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002F1.2.8...1.3.0\n\n本次发布主要聚焦于可靠性：针对 `efficient` 模型优化了重试时机，移除了未使用的调试断言，并进行了版本管理和例行维护更新，以提升系统在生产环境中的稳定性。\n\n## 亮点\n- 优化了 `efficient` 模型的重试时机，以减少瞬时失败。\n- 移除了未使用的调试断言，从而降低日志噪声并简化故障排查。\n- 提升了版本号并进行了例行维护，使发布管理更加清晰。\n\n## 升级说明\n- 预计无破坏性变更。请运行常规集成测试；重点关注在您负载下模型的重试行为。","2026-03-10T14:24:07",{"id":106,"version":107,"summary_zh":108,"released_at":109},135998,"1.2.8","## 🚀 v1.2.8 新增内容\n\n> 感谢贡献者 @alexeysviridov、@gurdasnijor 和 @cbbfcd  \n> **完整变更日志**：https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002F1.2.7...1.2.8\n\n### ✨ 新特性\n\n- **外部知识文档支持** - 新增支持将任何可用的外部文档集成到代码库中，从而增强知识管理和文档处理能力。\n- **本地文档集成优化** - 改进了本地文档的集成方式，提供分类支持和更先进的分块策略，以实现更好的组织与检索效果。\n- **文档链接与分析限制** - 增加了文档链接功能，并优化了分析限制，以提升代码分析的工作流效率。\n\n### 🐛 问题修复\n\n- 修复了缓存模块中的语法错误。\n- 移除了未使用的变量，使代码库更加整洁。\n\n### 📦 其他变更\n\n- 在 README 中添加了 Smithery 徽章。\n- 更新了文档内容。\n\n---\n\n## 🙏 特别致谢\n\n衷心感谢以下优秀的贡献者，正是他们的努力才使得本次发布成为可能：\n\n- **[@alexeysviridov](https:\u002F\u002Fgithub.com\u002Falexeysviridov)** - 在外部知识文档支持以及结合分块策略的本地文档集成方面做出了卓越贡献。\n- **[@gurdasnijor](https:\u002F\u002Fgithub.com\u002Fgurdasnijor)** - 负责添加了 Smithery 徽章。\n- **[@cbbfcd](https:\u002F\u002Fgithub.com\u002Fcbbfcd)** \u002F **bobihuang** - 解决了未使用变量的问题。\n\n同时，也感谢所有提供反馈、报告问题并帮助改进 Litho 的各位！🎉\n\n---\n\n**完整变更日志**：[v1.2.7...v1.2.8](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Flitho\u002Fcompare\u002Fv1.2.7...v1.2.8)","2026-02-01T08:14:42",{"id":111,"version":112,"summary_zh":113,"released_at":114},135999,"1.2.6","## 变更内容\n* 默认语言已更改为英语，并由 @alexeysviridov 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F68 中添加了 C# 代码支持。\n* 功能：由 @fishers66 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F69 中添加了 Swift 语言处理器。\n\n## 新贡献者\n* @alexeysviridov 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F68 中完成了首次贡献。\n* @fishers66 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F69 中完成了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002F1.2.5...1.2.6","2025-11-24T02:03:06",{"id":116,"version":117,"summary_zh":118,"released_at":119},136000,"1.2.5","## 变更内容\n* 由 @sopaco 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F67 中格式化代码，并在提示中添加语言说明。\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002F1.2.3...1.2.5","2025-11-19T02:03:11",{"id":121,"version":122,"summary_zh":123,"released_at":124},136001,"1.2.3","> 此次更新提升了整体性能和稳定性，包括针对 Ollama 提供商下各型号的兼容性优化，同时引入了更多代码用途，以提高面向后端项目的分析效率。\n\n## 变更内容\n* Ollama 提供商支持，由 @sopaco 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F50 中实现\n* 新增功能：默认排除 pnpm 锁文件，由 @cbbfcd 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F51 中实现\n* 新增功能：根据文件路径格式化代码并检测测试文件，由 @cbbfcd 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F54 中实现\n* 增加越南语支持：deepwiki-rs --target-language vi，由 @lethanhson9901 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F53 中实现\n* 添加 PHP 语言处理器，由 @LiuChinNan 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F58 中实现\n* 添加 Dao 和 Context 代码用途变体，由 @sopaco 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F59 中实现\n* 新增功能：添加 Ollama 结构化输出支持，由 @lethanhson9901 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F60 中实现\n\n## 新贡献者\n* @cbbfcd 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F51 中完成了首次贡献\n* @vovanduc 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F52 中完成了首次贡献\n* @lethanhson9901 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F53 中完成了首次贡献\n* @LiuChinNan 在 https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fpull\u002F58 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002F1.2.1...1.2.3","2025-11-17T06:30:32",{"id":126,"version":127,"summary_zh":128,"released_at":129},136002,"1.2.1","## 发行说明\n\n> 这些更改在保持 Litho 引擎自动生成架构文档的核心功能的同时，增强了其鲁棒性。本次发布改进了依赖项管理和配置处理。\n\n### 新特性\n新增对 Ollama 提供者的支持，实现本地运行大语言模型。使用 `--llm-provider` 参数即可启用 Ollama 集成。\n\n\u003Cimg width=\"840\" height=\"420\" alt=\"image\" src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F33692e48-0691-4127-b8f2-3521031eebf8\" \u002F>","2025-11-11T03:09:40",{"id":131,"version":132,"summary_zh":133,"released_at":134},136003,"1.2.0","## 发布说明\n\n> 这些更改在保持 Litho 引擎自动生成架构文档核心功能的同时，提升了其鲁棒性。本次发布优化了依赖管理和配置处理。\n\n### 新特性\n新增基于预设配置文件的启动方式，以简化参数输入和管理。  \n请参考上述链接中的 [**litho.toml**](https:\u002F\u002Fgist.github.com\u002Fsopaco\u002F8b6c90aaaed07c5b444cf16668735003)。将该配置文件放置于项目根目录下，即可直接运行 deepwiki-rs 程序。\n\n### 重要变更\n- **升级 rig-core 依赖**：从 0.22.0 版本升级至 0.23.1，以提升稳定性，并包含对 `CompletionError: ProviderError: {\"error\":{\"code\":null,\"param\":null,\"message\":\"[] is too short - 'tools'\",\"type\":\"invalid_request_error\"}}` 问题的修复（参见：https:\u002F\u002Fgithub.com\u002F0xPlaygrounds\u002Frig\u002Fpull\u002F1003）。\n- **恢复默认 LLM 提供商**：将配置中的默认 LLM 提供商由 Moonshot 恢复为 OpenAI。","2025-10-30T14:43:38",{"id":136,"version":137,"summary_zh":138,"released_at":139},136004,"1.1.7","# 1.1.7 版本发布说明\n\n## 变更\n- 大幅提升了与非 OpenAI 和 Anthropic 的大模型提供商的兼容性\n- 优化了 `Overview Editor Agent` 和 `Architecture Editor Agent` 的上下文处理机制，以提升文档生成性能\n- 增强了 AI 分析结果的缓存机制\n- 在边界分析器中新增了对 MVC 模式代码用途的支持\n- 在配置中新增了默认的排除目录：`__tests__`、`__mocks__`、`__fixtures__`\n- 更新了依赖项，其中包括将 rig-core 从 0.21.0 升级至 0.22.0","2025-10-16T13:38:23",{"id":141,"version":142,"summary_zh":143,"released_at":144},136005,"1.1.5","## deepwiki-rs v1.1.5 发行说明\n\n### 功能\n- **新增路由器边界支持**：系统现支持对路由器边界进行分析和文档化，提供全面的系统接口视图，包括 CLI 命令、API 端点和路由器路径。\n\n### 改进\n- **增强文档生成功能**：边界编辑器现可为路由器边界生成详细文档，包含路径描述、源位置和参数详情。\n- **更新日志记录**：边界分析器现提供更详尽的日志信息，包括已分析的路由器路径数量。\n\n### 变更\n- **版本号升级**：将软件包版本从 1.1.1 更新至 1.1.5。","2025-10-14T15:27:42",{"id":146,"version":147,"summary_zh":148,"released_at":149},136006,"1.1.0","## 🌍 Release v1.1.0 - Multi-Language Support\r\n\r\n### ✨ New Features\r\n\r\n**🌐 International Documentation Generation**\r\n- Added support for 7 languages: English, Japanese, Korean, German, French, and Russian, Chinese\r\n- Language-specific file naming and directory structures\r\n- Localized LLM prompts for better documentation quality in each language\r\n\r\n### 🚀 Usage\r\n\r\n```bash\r\n# Generate documentation in different languages\r\ndeepwiki-rs --target-language en     # English\r\ndeepwiki-rs --target-language de     # German\r\ndeepwiki-rs --target-language ja     # Japanese\r\ndeepwiki-rs --target-language zh     # Chinese\r\n```\r\n\r\nSupports language codes, full names, and native names (e.g., `chinese`, `中文`, `deutsch`).\r\n\r\n### 🔧 Technical Details\r\n\r\n- Type-safe language configuration with Rust enums\r\n- Seamless integration with existing CLI and config system\r\n- Backward compatible - defaults to English\r\n\r\n---\r\n\r\n**Full Changelog**: [v1.0.0...v1.1.0](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002Fv1.0.0...v1.1.0)\r\n","2025-10-13T11:17:32",{"id":151,"version":152,"summary_zh":153,"released_at":154},136007,"1.0.0","> A first Stable release of deepwiki-rs: an automated, CLI-first system for generating high‑fidelity architecture documentation from source code using multi‑agent pipelines and LLM-assisted semantic analysis.\r\n> Document reference: [Litho Overview](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fblob\u002Fmain\u002Fdocs\u002F1%E3%80%81%E9%A1%B9%E7%9B%AE%E6%A6%82%E8%BF%B0.md)\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fcompare\u002F0.9.11...1.0.0\r\n\r\n## Highlights\r\n- Automated end-to-end pipeline that turns code, configs and docs into structured artifacts (System Context, Domain Modules, Workflows, Component diagrams) and Markdown outputs.\r\n- Rust-native CLI tool focusing on reproducible, auditable document generation with strong caching to reduce LLM cost.\r\n- Multi-agent, pluginable architecture for language processors and LLM providers — designed for extensibility and CI\u002FCD integration.\r\n- Supports 10+ mainstream programming languages via a pluginized `LanguageProcessor` approach.\r\n\r\n## Key Features\r\n- Automatic generation of System Context, Domain Module, Workflow, Container and Component views in Markdown + Mermaid.\r\n- Multi-language static analysis (Rust, Python, Java, JavaScript\u002FTypeScript, Vue, React, Svelte, Kotlin, ...).\r\n- LLM integration with provider abstraction and ReAct multi‑round reasoning; model failover and token cost monitoring.\r\n- Prompt-hash based persistent caching to avoid repeated LLM calls and lower costs.\r\n- Shared `Memory` context for agent-to-agent data passing, producing reproducible outputs and enabling auditability.\r\n- Outputs are plain Markdown + Mermaid for easy version control and integration into documentation sites (MkDocs, Docusaurus, etc).\r\n\r\n## What's included in this release\r\n- CLI entry and configuration: `cli.rs`, `config.rs`\r\n- Memory & foundation: `memory\u002F`, `utils\u002F`\r\n- Preprocessing: `preprocess\u002F` (file scanning, language parsing, code analyzers)\r\n- Research \u002F analysis agents: `research\u002F` (SystemContextResearcher, DomainModulesDetector, etc.)\r\n- Document composition: `compose\u002F` (Overview, Architecture, Workflow, KeyModules editors)\r\n- Output plumbing: `outlet\u002F` (DiskOutlet, SummaryGenerator)\r\n- LLM client abstraction: `llm\u002Fclient\u002F`\r\n- Prompt cache: `cache\u002F`\r\n\r\n## Supported integrations\r\n- LLM providers (via abstraction): Moonshot, Mistral, OpenAI\u002FAnthropic\u002FGemini-style providers can be configured through `llm\u002Fclient\u002F`.\r\n- Outputs are plain Markdown + Mermaid diagrams, suitable for Git-based versioning and downstream documentation pipelines (CI\u002FCD, static site generators).\r\n- ","2025-10-06T02:40:11",{"id":156,"version":157,"summary_zh":158,"released_at":159},136008,"0.9.11","> A first public release of deepwiki-rs: an automated, CLI-first system for generating high‑fidelity architecture documentation from source code using multi‑agent pipelines and LLM-assisted semantic analysis.\r\n> Document reference: [Litho Overview](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fdeepwiki-rs\u002Fblob\u002Fmain\u002Fdocs\u002F1%E3%80%81%E9%A1%B9%E7%9B%AE%E6%A6%82%E8%BF%B0.md)\r\n\r\n## Highlights\r\n- Automated end-to-end pipeline that turns code, configs and docs into structured artifacts (System Context, Domain Modules, Workflows, Component diagrams) and Markdown outputs.\r\n- Rust-native CLI tool focusing on reproducible, auditable document generation with strong caching to reduce LLM cost.\r\n- Multi-agent, pluginable architecture for language processors and LLM providers — designed for extensibility and CI\u002FCD integration.\r\n- Supports 10+ mainstream programming languages via a pluginized `LanguageProcessor` approach.\r\n\r\n## Key Features\r\n- Automatic generation of System Context, Domain Module, Workflow, Container and Component views in Markdown + Mermaid.\r\n- Multi-language static analysis (Rust, Python, Java, JavaScript\u002FTypeScript, Vue, React, Svelte, Kotlin, ...).\r\n- LLM integration with provider abstraction and ReAct multi‑round reasoning; model failover and token cost monitoring.\r\n- Prompt-hash based persistent caching to avoid repeated LLM calls and lower costs.\r\n- Shared `Memory` context for agent-to-agent data passing, producing reproducible outputs and enabling auditability.\r\n- Outputs are plain Markdown + Mermaid for easy version control and integration into documentation sites (MkDocs, Docusaurus, etc).\r\n\r\n## What's included in this release\r\n- CLI entry and configuration: `cli.rs`, `config.rs`\r\n- Memory & foundation: `memory\u002F`, `utils\u002F`\r\n- Preprocessing: `preprocess\u002F` (file scanning, language parsing, code analyzers)\r\n- Research \u002F analysis agents: `research\u002F` (SystemContextResearcher, DomainModulesDetector, etc.)\r\n- Document composition: `compose\u002F` (Overview, Architecture, Workflow, KeyModules editors)\r\n- Output plumbing: `outlet\u002F` (DiskOutlet, SummaryGenerator)\r\n- LLM client abstraction: `llm\u002Fclient\u002F`\r\n- Prompt cache: `cache\u002F`\r\n\r\n## Supported integrations\r\n- LLM providers (via abstraction): Moonshot, Mistral, OpenAI\u002FAnthropic\u002FGemini-style providers can be configured through `llm\u002Fclient\u002F`.\r\n- Outputs are plain Markdown + Mermaid diagrams, suitable for Git-based versioning and downstream documentation pipelines (CI\u002FCD, static site generators).\r\n- ","2025-10-05T13:19:42",[161,173,181,189,197,206],{"id":162,"name":163,"github_repo":164,"description_zh":165,"stars":166,"difficulty_score":167,"last_commit_at":168,"category_tags":169,"status":62},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[170,50,171,172],"Agent","图像","数据工具",{"id":174,"name":175,"github_repo":176,"description_zh":177,"stars":178,"difficulty_score":167,"last_commit_at":179,"category_tags":180,"status":62},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[50,171,170],{"id":182,"name":183,"github_repo":184,"description_zh":185,"stars":186,"difficulty_score":40,"last_commit_at":187,"category_tags":188,"status":62},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",142651,"2026-04-06T23:34:12",[50,170,51],{"id":190,"name":191,"github_repo":192,"description_zh":193,"stars":194,"difficulty_score":40,"last_commit_at":195,"category_tags":196,"status":62},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107888,"2026-04-06T11:32:50",[50,171,170],{"id":198,"name":199,"github_repo":200,"description_zh":201,"stars":202,"difficulty_score":40,"last_commit_at":203,"category_tags":204,"status":62},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[205,50],"插件",{"id":207,"name":208,"github_repo":209,"description_zh":210,"stars":211,"difficulty_score":167,"last_commit_at":212,"category_tags":213,"status":62},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,"2026-04-06T11:19:32",[51,171,170,50]]