Autodesk MCP Platform
Lets any MCP-compatible AI assistant read and edit Autodesk Civil 3D drawings through tools for alignments, surfaces, corridors, pipe networks, quantity takeoff, and cut/fill, using a local bridge plugin and named pipes.
README
Autodesk MCP Platform
Let any MCP-compatible AI assistant read and edit Autodesk Civil 3D drawings through tools - alignments, surfaces, corridors, pipe networks, quantity takeoff, cut/fill and more.
AI client (Claude Desktop, Cursor, VS Code, Cline, ...)
| Model Context Protocol (stdio)
v
Autodesk.MCP.Server (TypeScript / Node.js - product-agnostic MCP server)
| JSON-RPC 2.0 over a local Windows named pipe
v
Civil3D.Bridge (C# / .NET 8 plugin loaded inside Civil 3D)
| Autodesk .NET API (in-process)
v
Autodesk Civil 3D 2025 / 2026
- No network: everything runs on local named pipes, ACL'd to the current user.
- Dynamic discovery: the server learns the entire tool catalog from the bridge at handshake time - no tool is hardcoded, and future products plug in with zero server changes.
- Safe editing: editing tools run in transactions and require confirmation.
Quick start
- Install the bridge: copy the
Civil3D.Bridge.Bundlefolder from a release into%APPDATA%\Autodesk\ApplicationPlugins\and restart Civil 3D. - Install the server:
npm install -g autodesk-mcp-server(ornpx -y autodesk-mcp-server). - Point your AI client at the server (ready-made configs in
examples/clients/).
See docs/QuickStart.md and docs/Installation.md for the full walkthrough.
Documentation
| Guide | Contents |
|---|---|
docs/QuickStart.md |
5-minute setup + first prompts |
docs/Installation.md |
bridge bundle + npm install, upgrade, uninstall |
docs/Configuration.md |
server/bridge options, env vars, logging, multi-bridge |
docs/Troubleshooting.md |
diagnosis by symptom |
docs/DeveloperGuide.md |
build, test, benchmark, release from source |
docs/ReleaseProcess.md |
versioning, tags, release pipeline |
docs/Compatibility.md |
supported products/clients/protocol guarantees |
docs/RELEASE-VALIDATION.md |
acceptance checklist before announcing |
docs/PERFORMANCE-BENCHMARKS.md |
measured metrics + how to run them |
docs/ARCHITECTURE.md |
full design, ADRs, sequence diagrams |
docs/FAQ.md |
frequently asked questions |
Examples
examples/clients/- Claude Desktop, VS Code, Cursor, Cline configurationsexamples/config/- server/bridge configs, env vars, logging, multi-bridgeexamples/prompts/- typical prompts by workflowexamples/workflows/- end-to-end session transcriptsexamples/json-rpc/- real wire messages (NDJSON over the pipe)
Development
npm run quality # full local quality gate (mirrors CI)
See docs/DeveloperGuide.md and docs/Contributing.md.
Releases
Version 1.0.0 (GA) with the Phase 9 production-hardening work complete
(see docs/PRODUCTION-HARDENING.md and docs/RELEASE-1.0.0-FINAL.md). Tags follow
v<semver>; see docs/ReleaseProcess.md.
License
MIT - see LICENSE.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。