Project Marvin

Appearance

Calendar Sync Landscape ​

Problem statement ​

You need event mirroring across multiple calendars, including Microsoft 365 / Outlook and Apple Calendar, with destination events marked private or reduced to blocker-level detail.

Typical user stories:

  • Accept a work meeting in Microsoft 365 and have it auto-block personal and contracting calendars.
  • Create a contracting event and have it auto-block work and personal calendars.
  • Keep everything synchronized without manually duplicating events.

Constraints that matter ​

  • Microsoft 365 supports rich API access through Microsoft Graph.
  • Apple Calendar usually enters the picture through CalDAV or iCloud app authorization.
  • Cross-provider sync gets difficult around recurrence, updates, deletes, and loop prevention.
  • Privacy rules vary by provider and connector.

Viable implementation paths ​

1. Power Automate ​

Strong fit when:

  • most calendars are Outlook or Microsoft 365
  • you want a low-code starting point
  • you accept slower runs and more operational friction

Weak fit when:

  • you need Apple Calendar as a first-class writable target
  • you want reproducible infrastructure-as-code
  • you want strong control over loops, retries, and state

2. Custom sync service ​

Strong fit when:

  • you need durable automation
  • you want scripted deployment
  • you need precise privacy policies
  • you need Apple via CalDAV and Microsoft via Graph in one engine

Weak fit when:

  • you want the lowest possible implementation effort right now

3. Existing open-source sync product ​

Strong fit when:

  • you want to stand something up quickly
  • the tool already supports the providers you need
  • you can accept the tool's event model and operational shape

Weak fit when:

  • you need very specific mirror semantics
  • you need narrow control over metadata or tenant policy
  1. MShekow/outlook-calendar-sync as a reference for the Power Automate algorithm and flow packaging.
  2. ridafkih/keeper.sh as the most relevant open-source example of a broader multi-provider sync engine.
  3. A custom Graph plus CalDAV service if your priority is long-term ownership and scripted deployment.

Decision ​

Current recommendation:

  • Use Power Automate only as a short-term proof of concept for M365-to-M365.
  • For the real Project Marvin target, prefer either:
    • adapting an existing open-source multi-provider sync engine, or
    • building a narrow custom service around Microsoft Graph plus CalDAV.