Sync MCP Server

Install Sync MCP server

claude mcp add --transport http sync https://sync.lucie.network/api/mcp

{
  "mcpServers": {
    "sync": {
      "url": "https://sync.lucie.network/api/mcp"
    }
  }
}

{
  "mcpServers": {
    "sync": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://sync.lucie.network/api/mcp"
      ]
    }
  }
}

{
  "servers": {
    "sync": {
      "type": "http",
      "url": "https://sync.lucie.network/api/mcp"
    }
  }
}

{
  "mcpServers": {
    "sync": {
      "serverUrl": "https://sync.lucie.network/api/mcp"
    }
  }
}

Context tax

Connecting this server adds ~1,911 tokens of tool definitions to every model turn (moderate, 5 tools). Measured 2026-10-19 via its hosted endpoint; o200k tokenizer, worst case for clients that load all tools up front. How we measure

Input prices as of 2026-10-06. Excludes your own prompt and output.

Tools (5)

• book_meeting499 tok

  Book a confirmed meeting on a Lucie Sync host's calendar at a specific start time. This is a real write: it creates a calendar event and emails both the host and the booker. Always call check_availability first and use one of the open slots it returns. Before calling, (1) show the booker the specific date/time you're about to book — in their own time zone — and get their explicit confirmation, and (2) collect their name, email, AND a brief reason for the meeting (ask for the reason if not given; it's required and shared with the host).

• book_group_meeting392 tok

  Confirm a group meeting at a specific time across several Lucie hosts (by @handle), creating ONE shared Google Calendar event with everyone invited. YOU (the connector's owner) are the organizer and host the event on your calendar — no organizer handle needed. This is a real write: it re-verifies every participant (you included) is free at that instant and refuses if anyone isn't — so it never double-books. Always call find_group_times first and copy an exact common-slot start. Before calling, confirm the time (in each person's zone), the participants, the title, and the reason. Requires you t

• create_group_event386 tok

  Create a Lucie group event (a poll) for several participants, pre-filled with candidate times when everyone is free (the intersection of their availability). Returns a vote link to share: participants pick which proposed times work, and people WITHOUT a Lucie handle can vote manually too. This is a write (creates a poll) but sends no emails and touches no calendars. Confirm the title, participants, and duration with the organizer before creating. Note: agent group scheduling is a Sync Pro feature — the organizer's email must be on Sync Pro (or a Lucie Membership).

• check_availability344 tok

  Return the open meeting slots a Lucie Sync host offers, plus their booking rules (durations offered, minimum notice, how far ahead you can book, buffer between meetings). Returns only bookable slots and policy — never the host's raw calendar. Use this to find a time before proposing one to the user. Slots are absolute instants (ISO-8601 UTC). Pass the booker's viewerTimeZone so times display in THEIR local zone — the booker should never have to do time-zone math.

• find_group_times288 tok

  Given several Lucie hosts (by @handle), return the meeting times when ALL of them are free — the intersection of their availability — for a meeting of the given length. Read-only: never books, never exposes anyone's raw calendar (only the shared open slots). Use it to propose a group time. Times are absolute instants; pass viewerTimeZone so they show in the organizer's local zone.

Freshness

Active — last maintenance signal 2d ago. The newest of the signals below sets the band.

1. Last commit (default branch)

   no data · GitHub

2. Latest release

   no data · GitHub

3. Package published

   no data · npm/PyPI

4. Registry entry updated

   2026-10-17 · 2d ago · official registry · v1.0.0

How it compares in Calendar & Scheduling

Against the most-starred calendar & scheduling servers we index. Token cost is per model turn, measured from each server's live tool list.