11fgb-cleaning-tools-mcp
Provides deterministic cleaning cost, time, crew, and chemical usage estimates for homes and offices via MCP and HTTP endpoints.
README
11FGB Cleaning Tools MCP
Free deterministic cleaning calculators for AI assistants, agents, and web applications. The hosted server returns structured estimates, plain-text fallback answers, and optional MCP Apps visualizations.
Production endpoint: https://mcp.11fgb.com/mcp
Documentation and interactive calculators: 11fgb.com/developers/mcp
Tools
| Tool | Intended use |
|---|---|
estimate_home_cleaning_cost |
Home and apartment price, labor, crew, and duration planning |
estimate_office_cleaning_cost |
Per-visit and monthly office cleaning budgets |
estimate_cleaning_time_and_crew |
Workload, visit duration, and crew sizing |
calculate_cleaning_chemical_usage |
Ready-to-use solution, concentrate, water, and generic package quantities |
All results are planning estimates, not binding quotes. Product labels and Safety Data Sheets override chemical calculations. Never mix cleaning chemicals unless the manufacturer explicitly instructs it.
Connect
Use the remote Streamable HTTP endpoint in any compatible client:
{
"mcpServers": {
"11fgb-cleaning-tools": {
"type": "http",
"url": "https://mcp.11fgb.com/mcp"
}
}
}
No account or API key is required.
Example
Ask an assistant with the MCP connected:
Estimate a standard one-time cleaning for a 1,900 sq ft, three-bedroom, two-bath house in ZIP 78704 with one dog.
The result includes a price range, labor hours, recommended crew, expected duration, assumptions, warnings, and a 30-day visual report URL.
Run locally
Requirements: Node.js 20+.
npm install
npm run dev
The development server listens on http://localhost:3400. Run all checks with:
npm run check
docker build -t 11fgb-cleaning-tools .
Configuration
Copy .env.example into your own environment. Production requires a random 32-byte REPORT_TOKEN_KEY encoded as base64. The optional private rate feed can override the bundled national and Central Texas demonstration rate cards.
The report token contains only calculator inputs accepted by the public schemas. Exact street addresses are not accepted. Tokens are AES-256-GCM encrypted, expire after 30 days, and are not persisted as user records.
HTTP endpoints
| Endpoint | Purpose |
|---|---|
POST /mcp |
MCP Streamable HTTP transport |
POST /api/v1/calculations/{calculator} |
Web calculator adapter |
GET /api/v1/reports/{token} |
Recompute an anonymous visual report |
GET /health |
Container health |
GET /.well-known/mcp/server-card.json |
Registry scanner fallback |
POST /api/v1/events |
Anonymous aggregate report interaction event |
Methodology and data
The calculation engine is deterministic and versioned. Public fallback rates live in data/rates.json; hosted production may supply fresher rate cards through CALCULATOR_DATA_URL. A result always identifies its methodology and data version and distinguishes a verified Central Texas market from a national planning fallback.
Privacy and security
- No user account is required.
- The server does not request or accept exact street addresses.
- Full chat prompts are not stored.
- Rate limits apply to anonymous calculations.
- Report links are encrypted and expire after 30 days.
- Security reports: 11fgb.com/contact-us.
License
Production deployment
The production workflow deploys an exact commit SHA to /opt/11fgb-mcp. It builds a tagged image, starts an isolated candidate on port 3401, verifies health and an MCP initialize handshake, switches the Compose service on port 3400, and retains the previous tagged image for rollback. The container is limited to 512 MB RAM and log/image rotation is bounded.
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。