Why I Built ElkDesk

Five products. Forty emails a day. Zero time to actually build.

I was drowning in Gmail, hunting for buried support threads instead of shipping features. If you're an indie hacker facing the same choice—stick with Gmail chaos or pay $100+/month for bloated enterprise software—this is why I built something simpler.

The Gmail Years

For years, I ran a blogging platform. Support emails came in, I replied through Gmail, everything worked. A few emails a day, manageable threads, happy customers.

Then I launched a few new apps. They grew. Fast.

Suddenly I was drowning. Support emails scattered across Gmail for five different products. Threads buried, follow-ups missed, customers waiting. The "priority inbox" decided a newsletter was more important than a bug report about payment failing.

I calculated it once: 40 emails × 5 minutes each = 3+ hours daily just on support. That's 15 hours a week. An entire dev day, gone.

The Band-Aid Solutions

I built a help center. "This will reduce support volume," I thought.

It didn't. Users still emailed the same five questions, even though the answers were right there in the docs. Nobody reads documentation before emailing.

Then Gmail added AI reply suggestions. "Finally," I thought, "AI that helps."

It was useless. Generic responses that had no idea what my products did. "Thank you for contacting us. We'll look into this." Great. Super helpful.

I needed AI that actually knew my products. That learned from how I answered questions. That got smarter with every reply I sent.

Gmail couldn't do that.

The Zendesk Experience

"Okay," I thought, "time for a real solution."

I signed up for Zendesk—the "simple helpdesk alternative" everyone recommends. The pricing page said "$25/month." Reasonable.

Two hours later, I was still in the admin panel configuring triggers and views. Hadn't even connected my email yet.

The interface had fifteen sections: Triggers, Automations, Macros, Views, SLAs, Schedules, Skills-Based Routing. I just wanted to reply to emails, not build NASA mission control.

Features I Didn't Need

• SLA management: I'm one person. No contractual service agreements needed.
• Workforce scheduling: There's no workforce. It's me.
• Skills-based routing: I'm routing tickets to myself.
• Customer satisfaction surveys with NPS tracking: Overkill for 20 emails/day.
• Talk, Chat, Guide, Explore, Gather, Sell: What even are these?

Then I realized: the $25 plan is just basic email ticketing. Want AI suggestions? That jumps to $69/month for features that should be included.

And I had multiple projects. Managing up to 5 separate products? That's $149/seat. Most are just me, but one has a co-founder who handles support. Wanted to give him access? Double it. Now I'm at $298/month for two people.

The complexity was the real killer though. After two hours of configuring triggers and views, I still hadn't replied to a single email. The tool designed for "customer support" needed its own support team to operate.

Every tool I tried followed the same pattern: start simple, chase enterprise customers, get bloated, get expensive. Solo operators like me subsidize features I'll never use.

The "Screw It, I'm Building This" Moment

I was comparing pricing pages for the fifth tool in two days. Each had massive feature comparison tables - 50 rows of checkmarks. I only cared about 8 of those rows.

I sketched the simplest possible support tool on paper in 10 minutes:

• Ticket list on the left
• Conversation on the right
• Fast, clean, done

That's when I knew: the problem isn't that building a support tool is hard. The problem is years of accumulated enterprise feature requests.

Why "ElkDesk"?

Honestly? The domain was available. And ElkDesk sounds like HelpDesk.

But the more I built it, the more it fit.

Elk aren't apex predators trying to dominate the forest. They're grounded. Strong enough to handle harsh winters, but not complicated about it. They're just... there. Doing their thing. Part of the natural landscape.

That's what support software should be.

Not some aggressive enterprise beast that takes over your entire workflow. Just capable, reliable, grounded. Handles the job without making it complicated.

Elk don't overcomplicate survival. ElkDesk doesn't overcomplicate support.

What I Wanted Different

1. Setup in 5 minutes - not 10 hours
2. Fast UI - loads instantly, not in 3-4 seconds
3. Honest pricing - seats included, no hidden per-user fees. One of my apps has a co-founder who handles support, but the rest I built solo. I needed him to access just his product's inbox without sharing my login or paying for an extra seat
4. AI that learns your product - knowledge base that grows automatically as you reply
5. Clean multi-product support - separate inboxes, no routing rules

The Knowledge Base That Writes Itself

This became the feature I'm most proud of.

When you send a reply in ElkDesk, the AI analyzes it. If it detects something new - a question you haven't answered before, a product detail worth saving - then it asks: "Should this go in the knowledge base?"

One click, and it extracts the answer, categorizes it, and adds it to your product's knowledge. It doesn't ask every time. Just when it matters.

Next time a customer asks the same question, the AI already knows the answer. It suggests a reply based on how you answered it before - in your voice, with your product details.

The knowledge base is just a markdown file. Want to kickstart it? I use Claude Code to analyze my project and populate the initial knowledge. Then it grows automatically from real support conversations.

After two weeks of using ElkDesk, I had 47 answers in my knowledge base. I wrote maybe 10 initially. The rest came from just doing support.

That's the difference between AI that helps and AI that knows.

The Core Philosophy

Support is a feature, not a department.

For indie hackers and small teams, support should take 30 minutes of your day - not become a full-time job managing a support system.

Features I Intentionally Left Out

• SLAs: If you need contractual SLAs, you need a different tool
• Complex ticket forms: An email is fine
• Workforce management: No workforce to manage
• Phone/chat integration: Email first (maybe chat later, if it stays simple)
• Community forums: That's a different product
• Detailed analytics dashboards: See patterns in your inbox, no charts needed

The goal isn't to be "Zendesk for small teams." It's to be the anti-Zendesk. Do five things well instead of fifty things adequately.

ElkDesk Is Live Right Now

I'm not asking you to take this on faith. ElkDesk is handling support for my own products as you read this. Five apps, hundreds of emails a month, real customers with real problems.

If it breaks, my customers wait. If the AI suggests bad replies, I'm wasting time instead of saving it. I'm using ElkDesk for my own revenue—real customers, real stakes. That's the only accountability that matters.

The Promise

I built this because I was frustrated. I'm keeping it simple because the moment it gets complicated, I'll be frustrated using my own product.

Will ElkDesk work for a 500-person support team? No. And I'm fine with that.

Trying to serve enterprise is exactly how every other tool became unusable for people like us.