Building a Calendar and Scheduling App with Harper

Building calendar features turns into a mess fast. What starts as "just store some events" becomes Redis for caching, Kafka for real-time updates, and multiple services just to check if someone's free for a meeting.

‍

You end up debugging why calendars aren't syncing instead of building the scheduling features users actually want.

‍

Harper changes this. Everything runs in one process: database, cache, real-time updates, conflict detection. No network calls between services, no complex synchronization.

The Traditional Calendar Nightmare

Calendar systems get complex because of coordination between services. Your availability service needs to talk to your event service. Notifications need to sync with conflict detection. Time zones need to work across everything.

A simple "meeting time changed" notification becomes a distributed systems problem. Database updates, cache invalidation, real-time pushes, and hoping nothing fails halfway through.

Setting Up Your Calendar Schema

Harper uses GraphQL to define your data model and automatically generates REST endpoints:

type User @table @export {
  id: ID @primaryKey
  email: String @indexed
  name: String
  timezone: String
  createdAt: String @createdTime
}

type Event @table @export {
  id: ID @primaryKey
  title: String @indexed
  startTime: String @indexed
  endTime: String @indexed
  organizerId: ID @indexed
  organizer: User @relationship(from: organizerId)
  attendeeIds: [ID] @indexed
  attendees: [User] @relationship(from: attendeeIds)
  status: String @indexed
  createdAt: String @createdTime
}

type Availability @table @export {
  id: ID @primaryKey
  userId: ID @indexed
  dayOfWeek: Int @indexed
  startTime: String @indexed
  endTime: String @indexed
  isAvailable: Boolean @indexed
}

The @export directive creates REST endpoints at /User/, /Event/, and /Availability/. The @relationship directive handles joins automatically.

Simple Scheduling Operations

Create a meeting:

POST /Event/
Content-Type: application/json

{
  "title": "Product Review",
  "startTime": "2024-08-15T14:00:00Z",
  "endTime": "2024-08-15T15:00:00Z",
  "organizerId": "user123",
  "attendeeIds": ["user456", "user789"]
}

Get someone's calendar:

GET /Event/?organizerId=user123&startTime=gte=2024-08-01&sort=startTime

Check for conflicts:

GET /Event/?attendeeIds=user123&startTime=lt=2024-08-15T15:00:00Z&endTime=gt=2024-08-15T14:00:00Z

Everything happens in the same process. No network latency, no coordination complexity.

Real-Time Updates

Harper has WebSockets built in:

ws://localhost:9926/Event/user123

When someone accepts a meeting, all attendees see the update instantly. No message brokers needed.

Time Zone Handling

Store everything in UTC, handle conversion at query time:

GET /Event/?organizerId=user123&timezone=America/New_York&startTime=gte=2024-08-15

No separate timezone service, no complex conversion logic.

Why This Works

Harper's single-process architecture eliminates the coordination problems that make calendar systems complex. You get faster response times, automatic conflict detection, and real-time sync without managing multiple services.

You focus on building scheduling features instead of debugging distributed systems.

‍

‍

‍