Hetk

Hetk syncs your calendars across Google, Outlook, and iCloud. April was features. May was paying for them.

I'll be honest, I shipped almost nothing this month that a user would notice. I spent it down in the sync engine, and it was probably the most important month since I started.

The incident that set the tone

Early in May, Microsoft Graph started returning 410s, which is its way of saying, "your sync state is stale, start over." Our handling turned that into a storm. The message backlog grew unbounded, climbing into the hundreds of thousands. The fix that was supposed to prevent this had looked fine in tests a few weeks earlier. It just didn't hold in production.

That pushed me to ask something I'd been avoiding: if a sync ever loops, what actually stops it? Nothing did. So I built a Loop Breaker. When it spots a relationship bouncing the same events back and forth, it pauses that one relationship instead of letting the whole queue melt. I ran it in observe-only mode for a few days, then armed it once I trusted its flagging.

iCloud was the hard child

Apple and CalDAV ate a big chunk of the month. There are no real webhooks, no quota headers, the throttling is undocumented, and the API breaks in ways Google and Microsoft don't. I added backoff for 429s and 503s, fixed recurring iCloud events that only synced their first instance, prevented privacy settings from being quietly dropped on CalDAV targets, and chased down a batch-write failure that only showed up at scale against real iCloud. Still not done, but a lot steadier than it was on the 1st.

Testing got serious

I built a cross-provider sync matrix that runs every source-to-target combination as a single suite, and it immediately revealed bugs I'd never have found by hand. To prevent iCloud testing from relying on flaky real accounts, I set up a local CalDAV server for the E2E suite. The whole thing runs in CI now and blocks a push if it goes red.

What's next

June goes back to features, starting with the one people keep emailing me about: filtering. Right now Hetk syncs everything in a calendar, all or nothing. Soon you'll be able to pick what crosses over. Sync only the events that match a rule, leave the rest behind, and rewrite titles per direction so a work meeting can show up on your personal calendar as "Busy" without dragging the details along. That's the bit I'm most excited about. It's a lot easier to build now that I trust the floor underneath it.

> "I have never found a Google/MS sync option that has worked effectively, and I'm beginning to think this one is no different. If this doesn't work, I want to back out completely."

That's a real user, day 22 of [Hetk](https://hetk.io)'s launch. I'm building it solo from Estonia.

He hit three things in one onboarding flow:

1. The wrong calendar showed up as "primary" at the sync step, even though he'd picked a different one earlier

2. Adding his own personal calendar triggered a paywall he shouldn't have hit

3. He paid anyway, and the action still failed afterward

I read his email, looked at the code, and found three separate bugs all hitting one new user at once. Each one is fixable in isolation. The problem was that fixing them in isolation would leave the original onboarding shape intact, and that shape itself was the issue.

So I rewrote it. The cheapest bug to fix is the one in a step that no longer exists.

Two things came out of onboarding entirely:

The calendar pre-selection step is gone. The wizard used to ask new users to pick which calendars they cared about, designate one as primary, and only then move on to creating a sync. Now there's no upfront calendar step.

The plan choice and credit card entry are also gone. Onboarding used to require both before you could reach the dashboard. Neither is required at signup anymore.

What the calendar step turned into: when you create your first sync, the wizard fetches calendars from the provider on demand. Searchable, grouped by account. There's no "primary" designation for the wizard to lose, because there's no designation step. You pick source and target at sync creation time, full stop.

The trial works differently now, too. Personal plan (3 sync pairs, 21 days) auto-starts at signup. No Stripe customer, no card. Card entry happens later, when one of three things triggers it: trial ending, you click upgrade, or you hit the Personal limit. That last one is a soft paywall, not a wizard block. You still review the sync pair you were trying to create. The Create button just changes to Upgrade to create. If the trial expires without payment, syncs pause, but the config sticks around. Pay later, and they resume instantly.

People sometimes ask about a free tier. At $2/mo for Personal, it would mostly cannibalize paying conversions rather than open a new channel. The no-card trial is handling acquisition for now. I'll think about it again if I reach 1,000 paying users, or if comparison sites start docking Hetk for not having one.

Two things I'd do differently next time:

I shipped without telemetry on dead-end states. Empty calendar dropdowns, paywall hits during sync creation, post-payment errors, and calendar fetches that return zero results. Any one of those firing on this user would have caught the bug before he sent the email. Wiring them up now.

I also got lucky. The user's report mapped cleanly to bugs already visible in the code. The 24-hour rewrite worked because the diagnosis was easy, not because I moved fast. Won't bet on that always being true.

Too early to say if this moved trial-to-paid conversion. I'll post numbers when I have them.

[Hetk](https://hetk.io) syncs calendars across Google, Outlook, and iCloud in both directions, with optional privacy controls that strip titles and attendees from synced events. $1 to $6 a month or $10 to $50 a year.

What onboarding step would you delete?

Hey IH 👋

I just launched Hetk — a service that syncs your events across Google Calendar, Microsoft Outlook, and Apple iCloud in real time.

The problem

If you use more than one calendar platform (work Outlook + personal Google, for example), you know the pain. You either manually copy events, set up fragile Zapier automations that lose half the event data, or accept that you'll occasionally double-book yourself.

I hit this myself and looked at existing solutions. Most are either Google-only, charge a monthly fee, or don't support bidirectional sync properly. None of them had the privacy controls I wanted — I don't need my coworkers seeing "Dentist appointment" on my work calendar.

What Hetk does

- Bidirectional sync between Google Calendar, Outlook, and iCloud (any combination)
- One-way sync if you want to mirror events in one direction
- Privacy controls — mark synced events as "Busy" and strip titles, descriptions, and attendees
- Real-time — uses webhooks, not polling, so changes show up in seconds
- 21-day free trial

Pricing

I went with annual pricing to keep it simple and cheap:

- Personal: $10/year (early adopter), $15/year regular
- Professional: $35/year (early adopter), $50/year regular

Annual billing only — no monthly subscriptions to worry about.

What's next

I'm looking for early users who deal with multi-calendar chaos daily. If that's you, I'd love to hear what your setup looks like and what matters most to you.

Check it out at hetk.io.

Happy to answer any questions about the product or the build.