mcp-meetsync
Integrates the MeetSync calendar negotiation API to enable AI agents to autonomously manage participants, find mutual availability, and handle meeting bookings. It exposes 19 tools for end-to-end scheduling workflows including participant preferences, proposals, and confirmations.
README
mcp-meetsync
MCP server for MeetSync — a calendar negotiation API built for AI agents.
Exposes all 19 MeetSync endpoints as MCP tools so any MCP-compatible LLM can autonomously find availability, propose meeting times, and confirm bookings — no human back-and-forth required.
Installation
npm install -g mcp-meetsync
Or run directly with npx:
npx mcp-meetsync
Configuration
The server reads two environment variables:
| Variable | Required | Default | Description |
|---|---|---|---|
MEETSYNC_API_URL |
No | http://localhost:3000 |
Base URL of your MeetSync API |
MEETSYNC_API_KEY |
Yes | — | API key sent in X-API-Key header |
Copy .env.example to .env and fill in your values, or pass them directly in the Claude tool definition.
Adding to Claude
Paste this snippet into your Claude desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"meetsync": {
"command": "npx",
"args": ["-y", "mcp-meetsync"],
"env": {
"MEETSYNC_API_URL": "https://api.yourmeetsync.com",
"MEETSYNC_API_KEY": "your-api-key-here"
}
}
}
}
After saving, restart Claude. The 19 MeetSync tools will appear in Claude's tool list.
Tools
All 19 MeetSync operationIds are exposed as tools. Tool names match operationIds exactly.
Participants (7 tools)
| Tool | When to use |
|---|---|
listParticipants |
Browse or search registered participants |
createParticipant |
Register a new person before they can be scheduled |
getParticipant |
Look up a specific participant's timezone/provider details |
updateParticipant |
Change a participant's name, email, timezone, or calendar setup |
deleteParticipant |
Permanently remove a participant (use force to cascade) |
getParticipantPreferences |
Read a participant's working hours and scheduling constraints |
setParticipantPreferences |
Define or replace working hours, blackout windows, and buffers |
Availability (2 tools)
| Tool | When to use |
|---|---|
getParticipantAvailability |
Find free windows for a single participant |
findMutualAvailability |
Find scored slots that work for all participants simultaneously |
Proposals (5 tools)
| Tool | When to use |
|---|---|
listProposals |
Browse pending or historical proposals |
createProposal |
Propose candidate time slots and send to participants for consensus |
getProposal |
Check proposal status and see who has responded |
cancelProposal |
Withdraw a pending proposal |
respondToProposal |
Record a participant's acceptance or rejection |
Bookings (5 tools)
| Tool | When to use |
|---|---|
listBookings |
Browse confirmed, cancelled, or rescheduled meetings |
createBooking |
Confirm a booking from an accepted proposal, or book directly |
getBooking |
Retrieve full details of a specific meeting |
rescheduleBooking |
Move a confirmed meeting to a new time |
cancelBooking |
Cancel a meeting permanently |
Example agent workflow
Here is a complete scheduling workflow an AI agent would follow using these tools:
1. createParticipant ← register alice@example.com
2. createParticipant ← register bob@example.com
3. setParticipantPreferences ← Alice: Mon–Fri 09–17, 15 min buffer
4. setParticipantPreferences ← Bob: Mon–Fri 10–18, no back-to-back
5. findMutualAvailability ← find 60-min slots next week for [alice, bob]
→ returns top 5 scored slots
6. createProposal ← "Budget Review" with top 3 slots, expires in 24h
→ proposal id: prop_abc123
7. respondToProposal ← alice accepts, prefers slot 1
8. respondToProposal ← bob accepts, prefers slot 1
→ proposal auto-transitions to "accepted", acceptedSlotId set
9. getProposal ← confirm acceptedSlotId
10. createBooking ← proposalId + slotId
→ booking confirmed, calendarEventIds written
Development
# Install dependencies
npm install
# Type-check without building
npm run typecheck
# Build to dist/
npm run build
# Run in dev mode (auto-reloads)
MEETSYNC_API_URL=http://localhost:3000 MEETSYNC_API_KEY=dev-key npm run dev
Architecture
src/
index.ts ← MCP Server, ListTools + CallTool handlers, stdio transport
client.ts ← Typed fetch wrapper: GET/POST/PUT/PATCH/DELETE + X-API-Key injection
tools/
participants.ts ← 7 participant tools + handlers
availability.ts ← 2 availability tools + handlers
proposals.ts ← 5 proposal tools + handlers
bookings.ts ← 5 booking tools + handlers
Each tool file exports:
- A
Tool[]array withname,description, andinputSchema(strict JSON Schema) - A
handle*Tool(name, args)async function that calls the MeetSync API viaclient
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。