ifs-mcp-server-local
Connect Claude to your IFS Cloud instance and interact with your ERP through natural conversation.
README
MCP Server for IFS Cloud
Connect Claude to your IFS Cloud instance and interact with your ERP through natural conversation.
The Idea
What if you could automate your IFS Cloud tasks without buying anything extra? No new subscriptions, no new platforms, no developers needed.
Install the extension, teach Claude some IFS skills, and then use it in your everyday life. Think of it as a coworker who never gets tired of the repetitive stuff.
The best part? You stay in control. You decide what the agent knows, how it behaves, and what it can do — the possibilities are only limited by your imagination.

[!NOTE] This product is not affiliated with IFS. It is a personal project by Damith Jinasena.
[!WARNING] This product is NOT tested in a production environment. It is still a proof of concept for creating a personal AI agent for IFS.
How It Works
Think of the IFS MCP server as a broker between Claude LLM and IFS Cloud. It can translate your questions into IFS API calls and bring the data from IFS. In order to understand what IFS Cloud API to call and to construct the message, it needs to know about the IFS endpoint metadata, what we call as skills.
What is a Skill?
Skill is a plain text .md file which contains instructions on IFS endpoints, their usage and the data structures. MCP server has instructions to create the file. You just need to tell it what you need!
There are several ways of creating a skill:
- From a projection name — Just give Claude the IFS projection name. It fetches the live OpenAPI spec and drafts the skill. Best for master data (customers, suppliers, parts).
- From a browser recording — Record your workflow in IFS using DevTools (F12 → Network → Save as HAR). Claude extracts the API calls, asks you to explain each step, and builds the skill. Best for multi-step transactional flows.
- Import a skill — If a friend or colleague has created an awesome skill, you can import that skill directly from a URL or local path. Have a look in the communty IFS MCP skills. You might find some useful skills.
Once you have the skill created, use it to perform any action supported by the skill.
Eg: If you create a skill from PartHandling projection, you can perform any operation related to parts just by telling claude!
Skills are the brain of the agent. More skills, more things you can do!
Share your skills in IFS MCP Skills so others can use them!
Quick Start
- Download the latest
ifs-mcp-server.mcpbfrom GitHub Releases and add it as an extension in Claude desktop. - Build your first skill.
- Now it's ready to use to perform your IFS tasks in Claude desktop!
See the installation guide → for complete instructions on IFS OAuth client setup, configuration, first authentication, and building your first skill.
Documentation
Getting Started
| Document | What's in it |
|---|---|
| Installation | Step-by-step setup: OAuth client, extension install, first authentication |
| Configuration | Skills directory setup and security notes |
Guides
| Document | What's in it |
|---|---|
| Skill Authoring | How to build, update, and share skills — HAR recordings and OpenAPI specs |
Reference
| Document | What's in it |
|---|---|
| Tools & Prompts | Full reference for all tools and prompts |
| Architecture | System design and component overview |
Community
| Document | What's in it |
|---|---|
| Contributing | How to contribute skills, report bugs, or develop the server |
| Security | Data handling, responsible use, and vulnerability reporting |
Built with the help of Claude Code. Shared with love for the IFS Community 💜
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。