让你的 AI 助手深入理解任何代码库
一个独立的 MCP 服务器，可将代码分析数据公开给 Gemini、Claude、Cursor、Windsurf、VS Code Copilot — 并在对话之间提供持久化的 AI 记忆。
Your Code → CodeAtlas Analyze → MCP Server → AI Understands Everything
│
10 powerful tools:
• Search functions & classes
• Trace feature flows
• Generate architecture diagrams
• Persistent memory across conversations使用 CodeAtlas 前： AI 盲目地进行 grep 搜索，在每次对话中都会忘记你的项目。 使用 CodeAtlas 后： AI 可以立即了解你的架构、依赖关系，并记住上下文。
Related MCP server: code-intel MCP Server
安装 CodeAtlas 扩展，然后：
Ctrl+Shift+P → CodeAtlas: Analyze Project{
"mcpServers": {
"codeatlas": {
"command": "npx",
"args": ["-y", "@giauphan/codeatlas-mcp"]
}
}
}{
"mcpServers": {
"codeatlas": {
"command": "npx",
"args": ["-y", "@giauphan/codeatlas-mcp"]
}
}
}{
"mcp": {
"servers": {
"codeatlas": {
"command": "npx",
"args": ["-y", "@giauphan/codeatlas-mcp"]
}
}
}
}{
"mcpServers": {
"codeatlas": {
"command": "npx",
"args": ["-y", "@giauphan/codeatlas-mcp"]
}
}
}claude mcp add codeatlas -- npx -y @giauphan/codeatlas-mcp{
"mcpServers": {
"codeatlas": {
"command": "npx",
"args": ["-y", "@giauphan/codeatlas-mcp"]
}
}
}完成！ 你的 AI 现在可以使用所有 10 个 CodeAtlas 工具了。
工具 | 描述 | 使用示例 |
| 列出所有已分析的项目 | “我有哪些项目？” |
| 获取模块、类、函数 | “向我展示所有类” |
| 导入/调用/包含关系 | “UserService 依赖什么？” |
| 代码质量与安全分析 | “有什么安全问题吗？” |
| 按名称模糊搜索实体 | “查找登录函数” |
| 特定文件中的所有实体 | “auth.ts 里有什么？” |
工具 | 描述 | 使用示例 |
| Mermaid 架构图（模块导入） | “向我展示系统架构” |
| Mermaid 执行流程图（调用链） | “支付功能是如何工作的？” |
工具 | 描述 | 使用示例 |
| 将上下文保存到 | “记住我们修改了什么” |
| 在代码库中追踪功能 | “哪些文件涉及身份验证？” |
AI 助手在对话之间会忘记一切。CodeAtlas 解决了这个问题：
Conversation 1 → AI analyzes code → sync_system_memory
│
.agents/memory/
├── system-map.md ← Architecture diagram
├── modules.json ← All entities
├── conventions.md ← Code patterns
├── business-rules.json ← Domain logic
├── feature-flows.json ← Feature traces
└── change-log.json ← Change history
│
Conversation 2 → AI reads memory → full context restored instantly ✨当你运行 Analyze Project 时，CodeAtlas 会为你的 AI IDE 自动创建规则文件：
生成的文件 | 适用对象 |
| 所有 AI 助手 |
| Cursor |
| Claude Code |
| Windsurf |
这些文件会告诉你的 AI：
在每次对话开始时读取 .agents/memory/
在进行更改前使用 MCP 工具
在完成更改后调用 sync_system_memory
语言 | 特性 |
TypeScript / JavaScript | 完整 AST：导入、类、函数、变量、调用、实现 |
Python | 类、函数、变量、导入、调用 |
PHP | 类、接口、特性 (traits)、枚举、函数、属性 |
Blade Templates |
|
npm install -g @giauphan/codeatlas-mcp然后在你的 MCP 配置中使用 "command": "codeatlas-mcp"（无需 args）。
变量 | 描述 |
| 强制指定项目目录 |
默认情况下，服务器会自动发现
~/下所有带有.codeatlas/analysis.json的项目。
git clone https://github.com/giauphan/codeatlas-mcp.git
cd codeatlas-mcp
npm install
npm run build
npm test # 6 tests
npm start # Start MCP serverCodeAtlas 扩展 — 带有交互式代码图的 VS Code 扩展
MCP 协议 — 模型上下文协议标准
MIT — 个人和商业用途免费。
Generate a Mermaid diagram showing the EXECUTION FLOW of a feature. Unlike generate_system_flow (which shows module imports), this traces the actual call chain: entry point → controller → service → model → database. Given a keyword, it finds all related functions and classes, then builds a flowchart or sequence diagram showing how they call each other at runtime. This is the best tool for understanding HOW a feature works step-by-step.
Inputs: keyword (required), project, diagramType, depth, +1 more.
Auto-generate a Mermaid flowchart diagram showing how modules, classes, and functions connect in the system. Returns a Mermaid diagram string that AI can read to understand the full system flow without reading every file.
Inputs: project, scope, feature, maxNodes.
Get import/call/containment/implements relationships between entities. Shows how modules, classes, and functions are connected.
Inputs: project, source, target, relationship, +1 more.
Get all entities (classes, functions, variables) defined in a specific file.
Inputs: filePath (required), project.
Get AI-generated code insights including refactoring suggestions, security issues, and maintainability analysis.
Inputs: no named parameters.
Get all modules, classes, functions, and variables in the analyzed project. Returns entity type, name, file path, and line number.
Inputs: project, type, limit.
List all projects that have been analyzed by CodeAtlas. Returns project names, paths, and last analysis time.
Inputs: no named parameters.
Search for functions, classes, modules, or variables by name. Supports fuzzy matching.
Inputs: query (required), project, type.
Create or update the .agents/memory/ folder with auto-generated system documentation. This folder serves as AI's 'long-term memory' — it persists between conversations. After calling this, AI in any future conversation can read these files to understand the full system flow without re-analyzing. Call this after completing any code changes.
Inputs: project, businessRule, changeDescription.
Trace the complete flow of a feature through the codebase. Given a keyword (e.g. 'login', 'payment', 'crawl'), finds all related files, classes, and functions, then orders them by dependency chain to show the execution flow. This helps AI understand which files to read when working on a feature.
Inputs: keyword (required), project, depth.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 10 tools
Most tools have distinct purposes. However, `generate_feature_flow_diagram` and `trace_feature_flow` both trace feature flow by keyword, differing only in output format (diagram vs. ordered list). Descriptions help differentiate, but an agent might still select the wrong one for a given need.
All tool names follow a consistent verb_noun pattern in snake_case (e.g., `generate_feature_flow_diagram`, `get_dependencies`, `list_projects`). No mixed conventions or vague verbs.
With 10 tools, the set covers the domain of code analysis and documentation well. Each tool serves a distinct, necessary function without redundancy or bloat.
The tool surface covers core CRUD-like operations for projects and entities, plus advanced analysis (flows, insights, memory sync). Minor gaps exist, such as no tool for editing entities or comparing versions, but core workflows are supported.
Search indexed code, trace dependencies, assess change impact, and recall repository memory.
The Cortex MCP server provides read-only access to real-time engineering context from the Cortex developer portal, allowing AI coding assistants to answer natural language questions about your organization's catalog (microservices, libraries, domains, teams, infrastructure), scorecards (engineering standards and best practices), initiatives (goals and deadlines), and Engineering Intelligence metrics. It includes tools for querying documentation, tracking personal entities, and accessing AI-assisted insights across the entire Cortex ecosystem.
Give your AI agent a persistent map of your project's structure, dependencies, and bugs.
Code intelligence for LLMs. Analyze, search, and retrieve code from any public git repository.
