atiendo24-mcp-server
MCP server that connects Claude Desktop to Atiendo24 platform, enabling management of appointments, customers, conversations, leads, analytics, and WhatsApp messaging through natural language.
README
atiendo24-mcp-server
MCP server for Atiendo24 — control your WhatsApp AI agent platform directly from Claude Desktop.
<img src="https://img.shields.io/badge/MCP-Compatible-blue?logo=anthropic" alt="MCP Compatible" />
What is this?
This is a Model Context Protocol (MCP) server that connects Claude Desktop to the Atiendo24 platform. It gives Claude access to 9 tools that let you manage appointments, customers, conversations, leads, analytics, and send WhatsApp messages — all through natural language.
You: "How many appointments do I have tomorrow?"
Claude: Uses get_appointments → "You have 3 appointments tomorrow: ..."
You: "Send a WhatsApp to +50688881234 saying the appointment is confirmed"
Claude: Uses send_whatsapp → "Message sent successfully."
Architecture
Claude Desktop (or any MCP client)
↓ MCP Protocol (stdio)
atiendo24-mcp-server (runs locally)
↓ ↓
Supabase DB WhatsApp Cloud API
(read/write) (send messages)
Tools
Read Tools
get_appointments
Fetches scheduled appointments with filtering options.
| Parameter | Type | Description |
|---|---|---|
filter |
"today" | "week" | "all" |
Time period (default: "today") |
client_id |
string |
Filter by business client ID |
status |
string |
Filter by status: confirmed, completed, cancelled, no-show |
Example prompts:
- "What appointments do I have today?"
- "Show me all cancelled appointments this week"
- "How many confirmed appointments are there?"
get_customers
Searches the customer database by name, phone, or email.
| Parameter | Type | Description |
|---|---|---|
search |
string |
Search by name, phone, or email |
client_id |
string |
Filter by business client ID |
limit |
number |
Max results (default: 20) |
Example prompts:
- "Find the customer with phone 8888-1234"
- "Search for Maria Lopez in the database"
- "How many customers do we have?"
get_conversations
Retrieves WhatsApp conversations with status filtering.
| Parameter | Type | Description |
|---|---|---|
status |
"active" | "escalated" | "resolved" |
Conversation status |
client_id |
string |
Filter by business client ID |
search |
string |
Search by customer name or phone |
limit |
number |
Max results (default: 20) |
Example prompts:
- "Are there any escalated conversations?"
- "Show me recent active conversations"
- "How many conversations were resolved today?"
get_leads
Fetches sales leads captured by the AI sales agent (Millie).
| Parameter | Type | Description |
|---|---|---|
status |
string |
Lead status: contacted, demo_scheduled, proposal_sent |
search |
string |
Search by name, phone, or business type |
limit |
number |
Max results (default: 30) |
Example prompts:
- "How many leads does Millie have this week?"
- "Show me leads with demos scheduled"
- "Find leads from dental clinics"
get_funnel
Returns the full sales conversion funnel with data quality metrics.
No parameters required.
Returns:
- Funnel stages: total leads → contacted → demo scheduled → proposal sent → closed clients
- Conversion rates between each stage
- Data quality: % of leads with valid phone, % with name
Example prompts:
- "How's the sales funnel looking?"
- "What's our lead-to-client conversion rate?"
- "What percentage of leads have valid contact info?"
get_analytics
Platform-wide metrics for a given time period.
| Parameter | Type | Description |
|---|---|---|
period |
7 | 30 | 90 |
Days to analyze (default: 7) |
client_id |
string |
Filter by business client ID |
Returns:
- Conversations: total, escalated, resolved, resolution rate, escalation rate
- Appointments: total, no-shows, cancelled, no-show rate
- Leads: total count
Example prompts:
- "Give me the metrics for the last 30 days"
- "What's the no-show rate this month?"
- "How many conversations were escalated last week?"
Write Tools
create_appointment
Creates a new appointment after checking availability.
| Parameter | Type | Required | Description |
|---|---|---|---|
client_id |
string |
Yes | Business client ID |
customer_phone |
string |
Yes | Customer phone number |
customer_name |
string |
Yes | Customer name |
date |
string |
Yes | Date in YYYY-MM-DD format |
time |
string |
Yes | Time in HH:MM format |
service |
string |
No | Service requested |
duration_minutes |
number |
No | Duration (default: 60) |
notes |
string |
No | Additional notes |
Example prompts:
- "Schedule an appointment for Juan Perez tomorrow at 10am"
- "Book a dental cleaning for Maria on 2026-05-01 at 14:00"
cancel_appointment
Cancels an existing appointment by ID.
| Parameter | Type | Required | Description |
|---|---|---|---|
appointment_id |
string |
Yes | Appointment ID to cancel |
client_id |
string |
No | Business client ID for verification |
Example prompts:
- "Cancel Juan's appointment"
- "Cancel appointment ID abc-123"
send_whatsapp
Sends a WhatsApp message from the Atiendo24 business number.
| Parameter | Type | Required | Description |
|---|---|---|---|
to |
string |
Yes | Destination phone (e.g., +50688881234 or 88881234) |
message |
string |
Yes | Message text |
Note: WhatsApp Cloud API requires the recipient to have messaged your number within the last 24 hours. Messages outside this window require pre-approved templates.
Example prompts:
- "Send a WhatsApp to +50688881234 saying the appointment is confirmed"
- "Message this prospect that I'll call them tomorrow"
Installation
Prerequisites
- Node.js >= 18
- A Supabase project with the Atiendo24 schema
- WhatsApp Cloud API credentials (for
send_whatsapp)
Setup
git clone https://github.com/edgarconejo/atiendo24-mcp-server.git
cd atiendo24-mcp-server
npm install
npm run build
Configure Claude Desktop
Add the following to your Claude Desktop config file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"atiendo24": {
"command": "node",
"args": ["/absolute/path/to/atiendo24-mcp-server/dist/index.js"],
"env": {
"SUPABASE_URL": "https://your-project.supabase.co",
"SUPABASE_SERVICE_ROLE_KEY": "your-service-role-key",
"WA_CLOUD_API_TOKEN": "your-whatsapp-cloud-api-token",
"WA_CLOUD_PHONE_NUMBER_ID": "your-phone-number-id"
}
}
}
}
Restart Claude Desktop after saving the config.
Environment Variables
| Variable | Required | Description |
|---|---|---|
SUPABASE_URL |
Yes | Your Supabase project URL |
SUPABASE_SERVICE_ROLE_KEY |
Yes | Supabase service role key (bypasses RLS) |
WA_CLOUD_API_TOKEN |
For send_whatsapp |
WhatsApp Cloud API permanent token |
WA_CLOUD_PHONE_NUMBER_ID |
For send_whatsapp |
WhatsApp Business phone number ID |
Verify Installation
After restarting Claude Desktop, click the + button in a new chat → Connectors — you should see "atiendo24" listed and enabled.
Try asking: "How many appointments do I have today?"
Tech Stack
- Runtime: Node.js
- MCP SDK: @modelcontextprotocol/sdk
- Database: Supabase (PostgreSQL)
- Messaging: WhatsApp Cloud API (Meta Graph API v21.0)
- Language: TypeScript
About Atiendo24
Atiendo24 is a multi-tenant SaaS platform that deploys AI-powered WhatsApp agents for businesses in Costa Rica. Each agent handles customer inquiries, schedules appointments, sends reminders, and processes payments — 24/7.
Built with Next.js, Supabase, Claude API, and WhatsApp Cloud API.
License
MIT
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。