ReAI MCP Server
MCP server for FSM report migration analysis, providing fuzzy search of report mappings, pattern registry, and columnar report generation.
README
ReAI MCP Server
MCP (Model Context Protocol) server that provides Kiro Web access to ReAI's core capabilities for FSM report migration analysis.
Tools
| Tool | Description |
|---|---|
reai_search_v10v11 |
Fuzzy search 967+ S3/Lawson to FSM report mappings with navigation paths |
reai_v10v11_modules |
List all FSM modules with report counts |
reai_v10v11_stats |
Database overview: totals, module breakdown |
reai_search_bre100 |
Fuzzy search 617+ report migration patterns from 11 clients |
reai_bre100_high_value |
Get standardization opportunities (multi-client patterns) |
reai_bre100_stats |
Pattern registry statistics by process and complexity |
reai_generate_columnar |
Generate Columnar LPL code package (.form, .list, .field, .csv) |
Installation
For Kiro Web (recommended)
Add to Kiro Web MCP settings:
{
"mcpServers": {
"reai": {
"command": "npx",
"args": ["-y", "github:van0219/reai-mcp#main"],
"env": {
"REAI_DATA_DIR": "/path/to/reai_web/data"
}
}
}
}
For Local Development
cd reai-mcp
npm install
npm start
Environment Variables
| Variable | Description | Default |
|---|---|---|
REAI_DATA_DIR |
Directory containing SQLite databases | ./data (bundled with package) |
REAI_V10V11_DB |
Full path to v10v11_mapping.db | $REAI_DATA_DIR/v10v11_mapping.db |
REAI_BRE100_DB |
Full path to bre100_registry.db | $REAI_DATA_DIR/bre100_registry.db |
Note: You typically don't need to set any env vars. The databases are bundled. Only set REAI_DATA_DIR if you want to point to a different copy (e.g., the Hub's reai_web/data/ for live writes).
Database Requirements
The server bundles two SQLite database files in its data/ folder — they ship with the package automatically:
-
v10v11_mapping.db (352KB) — S3/Lawson to FSM report mapping
LawsonReportMappingtable: id, report_code, report_name, status, module, descriptionFSMNavigationtable: id, report_id, navigation_path, path_order
-
bre100_registry.db (283KB) — BRE-100 pattern registry
patternstable: id, report_name, business_process, fsm_classification, navigation, future_state, tool_selection, timeline_estimate, complexity, comments, utilization_count, client_projects, created_at, updated_at
No additional setup needed — npx github:van0219/reai-mcp#main includes everything.
Updating the Databases
The databases grow as new clients are onboarded and patterns are discovered. To update:
From the Hub (Kiro IDE)
The Hub's native ReAI tool writes directly to reai_web/data/*.db via better-sqlite3. When patterns are added/edited in the Hub UI, those databases are the source of truth.
To sync updates to the MCP server:
- Copy the updated databases from the Hub repo to the MCP repo:
cp reai_web/data/v10v11_mapping.db reai-mcp/data/ cp reai_web/data/bre100_registry.db reai-mcp/data/ - Commit and push:
cd reai-mcp git add data/ git commit -m "Update databases (new patterns/mappings)" git push - New Kiro Web sessions auto-fetch the latest
#main— no config change needed.
From Kiro Web
Ask Kiro: "Copy the latest ReAI databases from the Hub repo to the reai-mcp repo and push them."
Kiro Web can clone both repos, copy the .db files from fsm_innovationhub_landingpage/reai_web/data/ to reai-mcp/data/, and push.
Database Locations Summary
| Context | Location | Writable |
|---|---|---|
| Hub (desktop) | reai_web/data/*.db |
Yes (via UI + IPC handlers) |
| MCP server (Kiro Web) | reai-mcp/data/*.db (bundled) |
Read-only (query only) |
| Override | Set REAI_DATA_DIR env var |
Depends on path |
Example Usage
Search V10-V11 Mappings
Use reai_search_v10v11 with query "GL Transaction" to find how S3 General Ledger reports map to FSM.
Search BRE-100 Patterns
Use reai_search_bre100 with query "asset depreciation" to find known migration patterns.
Generate Columnar Report
Use reai_generate_columnar with:
- report_name: "GLTransactionDetail"
- business_class: "GLTransactionDetail"
- fields: [{"name": "Company"}, {"name": "TransactionAmount"}, {"name": "PostingDate"}]
Author
FSM Innovation Hub — Infor Manila FSM Team
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。