Understand Anything MCP

MCP开发工具

提供方 uamcp · 包名 ua-mcp

开发工具 stdio · 本地进程 会话登录 社区 / 第三方 Claude DesktopClaudeCursorCodex
查看接入方式

A MCP server that empowers your AI agents to understand your entire project architecture, and a headless CI gateway to enforce architectural rules before code is merged. Prerequisite: ua-mcp is a lightweight reader tha

一句话选型结论

按资料字段整理

Understand Anything MCP 属于开发工具类 MCP 服务器,stdio · 本地进程,会话登录,可直接接入 Claude Desktop、Claude、Cursor 等,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类开发工具 传输方式stdio · 本地进程 鉴权会话登录 部署方式本地部署
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

开发工具
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

governance code analysis architecture developer tools impact analysis dependency graph ci cd code review

支持客户端

适用场景

governancecode analysisarchitecture

安装与配置

stdio · 本地进程 会话登录
客户端在本机拉起子进程,数据不出本机
npx -y ua-mcp

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 node,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
代码与仓库操作

在对话里直接查代码、提 issue、走 PR,省掉来回切工具。

本地命令与脚本

把构建、测试、格式化这类重复命令交给模型执行。

接口联调与测试

让模型调你的接口、跑一遍用例,快速定位问题。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 Claude Desktop、Claude、Cursor,换客户端不用重新研究配置。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 使用前要准备好会话登录对应的凭据,并按最小权限配置。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在开发工具场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 不方便管理密钥或第三方授权的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Understand Anything MCP当前 未标注 stdio · 本地进程 本地部署 有 未标注
Git 838e2ec1 12 stdio · 本地进程 本地部署 有 2026-05-17
Markitdown MCP 2c653a5f 1 stdio · 本地进程 本地部署 有 2026-05-17
JavaGuide 未标注 stdio · 本地进程 本地部署 有 2026-04-11
awesome-claude-skills 未标注 stdio · 本地进程 本地部署 有 2026-04-26
Context7 831acf40 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

开发工具
字段来自项目仓库与公开索引,中文说明由本站标注
提供方uamcp上游标注的维护方 uamcp
分类开发工具本站按上游分类展示
连接方式stdio · 本地进程客户端在本机拉起子进程,数据不出本机
鉴权方式会话登录远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时ua-mcp · node本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验未标注上游最后核验时间

详细介绍

9.6 KB
来自项目仓库与上游目录的原文,未做删改

Understand-Anything MCP Server

A Model Context Protocol (MCP) server that empowers your AI agents to understand your entire project architecture, and a headless CI gateway to enforce architectural rules before code is merged.

Quick Start

> [!NOTE]
> Prerequisite: ua-mcp is a lightweight reader that connects your AI assistant to your local Understand-Anything knowledge graph. It does not build the graph itself.
> You must install the core scanner from Egonex-AI/Understand-Anything and run /understand in your project to produce the .ua/knowledge-graph.json file. Commit this file to your repository before proceeding.

Get your AI assistant hooked up with architectural context in 60 seconds. ua_find_callers, ua_impact_analysis, and ua_precheck work immediately for free with no license key required!

Claude Desktop

Add this to your claude_desktop_config.json:```json
{
"mcpServers": {
"understand-anything": {
"command": "npx",
"args": ["-y", "ua-mcp"],
"env": {
"UA_PROJECT_PATH": "/absolute/path/to/your/project"
}
}
}
}


### Cursor
1. Go to **Settings > Features > MCP**.
2. Click **+ Add new MCP server**.
3. Name: `understand-anything`
4. Type: `command`
5. Command: `npx -y ua-mcp`
6. Add an environment variable: `UA_PROJECT_PATH = /absolute/path/to/your/project`

## How UA-MCP compares to Understand-Anything Tool

The free upstream Understand-Anything tool is responsible for generating the local knowledge graph and performing manual local analysis. **UA-MCP** sits on top of this graph to provide an automated **governance layer** and **safety net** for your codebase. While the free tool is a diagnostic scanner, UA-MCP actively blocks risky merges via its CI/branch-protection gate (`ua-ci`) and enforces your custom `.ua-rules.json` architectural boundaries.

## Privacy Policy
> [!IMPORTANT]
> **Link to Privacy Policy**
>
> **Data Processing Details:**
> - License keys and email addresses are securely stored for billing purposes.
> - **Purely local, no network calls:** `ua_status`, `ua_scan`, `ua_graph_summary`, `ua_explain`, `ua_onboarding_doc`.
> - **Sends graph data to the backend:** `ua_precheck` (Free and Pro), and all Pro-only tools (`ua_rules`, `ua_ci_check`, `ua_find_callers`, `ua_impact_analysis`, `ua_validate_graph`). When these tools are used, the full local graph object is sent to our backend for CI risk analysis, rule evaluation, and quota validation. All Pro-tier computation is handled remotely.
> - **No source code contents are transmitted**, only graph metadata (file paths and import relationships). All backend graph processing is done purely in-memory per-request and is never persisted.
>
> ### Security Considerations
> - `UA_API_URL` defaults to the official backend (`https://ua-mcp-backend.onrender.com`). Be extremely cautious if you change this to a custom endpoint, as the third-party backend will receive your full knowledge graph and source file paths.
> - **Self-Hosting:** If you are self-hosting the backend, note that a PostgreSQL `DATABASE_URL` is required in production. SQLite is not supported for production deployments.

## 🚀 Features

- **Branch Protection / CI Gate:** Automatically block high-risk PRs based on the codebase graph.
- **Architectural Rules:** Define custom `.ua-rules.json` to enforce boundaries.
- **Blast Radius Analysis:** Detect exactly which downstream files will break if a module is modified.
- **On-Demand Knowledge Graph:** Let Claude instantly query dependencies across massive codebases without filling its context window.

> [!TIP]
> **System Instruction Recommended:** For the best experience, add the following to your AI assistant's system prompt or custom instructions:
> *"Always call `ua_precheck` before modifying any file in this project."*

## How it works
1. Your AI agent decides it wants to modify a critical file (e.g., `src/auth.ts`).
2. The agent (following its system instructions) triggers `ua_precheck` before making the edit to run an **Architectural Safety Linter**.
3. The server analyzes the graph to determine the "blast radius" and checks it against your rules.
4. If the blast radius is too large or violates a rule, a **Safety Checkpoint** (Elicitation Prompt) interrupts the agent, asking for your explicit confirmation before proceeding.

## Configuration Rules (.ua-rules.json)
You can define specific boundaries in a `.ua-rules.json` file in the root of your workspace to dictate what the LLM is allowed to touch.```json
{
  // Understand-Anything Architectural Rules
  // Define constraints that agents and developers must respect.
  "rules": [
    {
      "id": "no-ui-db-import",
      "description": "UI layer must never import database layer directly",
      "from_pattern": "src/ui/**",
      "to_pattern": "src/db/**",
      "severity": "error"
    },
    {
      "id": "auth-required-for-payments",
      "description": "Payment modules must always be reachable from auth",
      "requires_path_through": "src/auth/**",
      "for_pattern": "src/payments/**",
      "severity": "error"
    }
  ]
}

🔒 Usage: CI/CD Branch Protection (Enforcement)

CI Gateway (ua-ci)

A companion CLI that runs locally in your GitHub Actions or GitLab CI. It parses your PR diff and compares it to the local graph.
- Free Tier: Evaluates blast-radius and logs the risk level.
- Pro Tier: Automatically blocks the merge if the risk is HIGH or violates architectural rules.

> [!IMPORTANT]
> The True Enforcement Backstop: While local agents rely on system instructions to run ua_precheck, the ua-ci command is designed to be your unbypassable safety net.
> By running ua-ci in your GitHub Actions and requiring it as a status check in GitHub Branch Protection, you ensure that no rogue edits can ever be merged into production without explicit approval.
>
> Note on Free Tier: If a valid Pro license key is not detected in the environment variables, the CI check will log a warning and silently pass (exit 0) so it does not block builds for non-paying users.```yaml

.github/workflows/ua-ci.yml

name: Understand-Anything CI Check
on:
pull_request:
branches: [ main ]

jobs:
ua-ci-check:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
with:
fetch-depth: 0

  - name: Generate PR Diff
    run: git diff origin/main...HEAD > pr.diff

  - name: Run UA Branch Protection
    env:
      UA_LICENSE_KEY: ${{ secrets.UA_LICENSE_KEY }}
    run: npx ua-ci --pr-diff=pr.diff

## 🛡️ Usage: Local Governance (Agents)

When connected to Claude Desktop or an MCP client, the following tools become available to the agent:

### Claude Desktop (`claude_desktop_config.json`)
Add the following to your Claude Desktop config file (usually `~/Library/Application Support/Claude/claude_desktop_config.json` on macOS or `%APPDATA%\Claude\claude_desktop_config.json` on Windows):```json
{
  "mcpServers": {
    "understand-anything": {
      "command": "npx",
      "args": ["-y", "ua-mcp"],
      "env": {
        "UA_PROJECT_PATH": "/path/to/your/project",
        "UA_LICENSE_KEY": "your_license_key_here"
      }
    }
  }
}

Cursor

  1. Go to Settings > Features > MCP.
  2. Click + Add new MCP server.
  3. Name: understand-anything
  4. Type: command
  5. Command: npx -y ua-mcp

Continue

Add to your config.json under mcpServers:json "understand-anything": { "command": "npx", "args": ["-y", "ua-mcp"] }

Available Tools & Tiering

The Understand-Anything MCP Server operates on a tiered licensing model.

Core Tools (Free Tier)

Available out of the box with no license required.
- ua_find_callers: Retrieves reverse dependencies up to 2 hops (Now unlimited and free!).
- ua_impact_analysis: Retrieves full transitive closure of reverse dependencies (Now unlimited and free!).
- ua_precheck: Pre-flight architectural risk check (10 checks/day, default critical-path rules only)
- ua_status: Returns MCP health status.
- ua_scan: Forces a re-scan of the workspace.
- ua_graph_summary: Returns aggregated node/edge statistics.
- ua_architecture_report: Groups files by top-level modules.
- ua_dependency_report: Identifies files with the most incoming dependencies (fan-in).
- ua_explain: Retrieves 1-hop dependencies for a specific file.
- ua_onboarding_doc: Generates onboarding context.

Premium Tools (Pro Tier)

Pro Tier — Architectural Enforcement & CI Safety

Team Use: One Pro license key can be shared across your entire team's CI pipelines and MCP configurations — no per-seat pricing.

  • ua_rules & ua_rules_check: Enforce custom .ua-rules.json boundaries. Evaluates constraints to ensure recent changes haven't introduced violations.
  • ua_ci_check & ua-ci: Block risky PRs in GitHub Actions before they reach production. Analyzes Git PR diffs for architectural impact.
  • ua_precheck: Unlimited pre-flight checks with configurable critical paths and .ua-rules.json enforcement
  • ua_validate_graph: Checks the knowledge graph schema for corruption.

Pricing

Tier Price Features
Free $0 forever Basic graph operations, local storage.
Pro $10/month OR $50 one-time Unlimited nodes, advanced graph analytics, rule enforcement, priority support. (Lifetime access limited availability)

Get your license key:
- Pro - Monthly ($10)
- Pro - Lifetime ($50)

Troubleshooting

  • Server fails to start: Ensure you have Node.js v18 or later installed.
  • License key error: Verify your key in the .env file or Claude config matches the one on your dashboard.
  • Path not found: Ensure UA_PROJECT_PATH is absolute or resolves correctly relative to where the server runs.

License

MIT License

常见问题

8 问
答案由本站依据条目字段整理,不引用上游原文
Understand Anything MCP 怎么安装?

在客户端按下面的安装命令接入:npx -y ua-mcp。对应的包名是 ua-mcp。装完在客户端配置里引用同一个包或端点。

Understand Anything MCP 怎么在 Claude Desktop、Claude、Cursor 里配置?

它的传输方式是stdio · 本地进程,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Understand Anything MCP 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Understand Anything MCP 需要 API Key 吗?

资料标注需要会话登录认证,使用前准备好对应凭据,并按最小权限配置。

Understand Anything MCP 支持哪些客户端?

资料标注支持 Claude Desktop、Claude、Cursor、Codex、Zed;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Understand Anything MCP 能提供哪些能力?

资料显示能力标签包括 governance、code analysis、architecture、developer tools、impact analysis、dependency graph。具体工具清单以官方文档为准。

Understand Anything MCP 需要什么运行环境?

运行时为 node;属于本地进程,需要在本机装好运行环境。

Understand Anything MCP 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。传输方式为stdio · 本地进程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 未标注。