The high after shipping

The hackathon ended and My Job Odyssey was live on the internet. A real product, with real users signing in, talking to Alex, building career memories, and watching their resumes update in real time. The early feedback was genuinely encouraging — people were using the weekly voice check-in not because I asked them to, but because it felt useful. One tester told me it was the first time anyone had asked them about their week at work. That hit different.

I was riding the wave. The Gemini Live voice pipeline was stable. Barge-in worked — users could interrupt Alex mid-sentence and the conversation would pivot naturally, just like talking to a real person. The career graph was growing. Sessions were getting smarter because the agent actually remembered context from weeks ago. For the first time, this thing I'd been building nights and weekends felt like it had legs.

But underneath the excitement, there was a question I couldn't ignore.

The cost fear

How much is this going to cost me?

I'd built an AI-powered platform that streams real-time bidirectional audio through the Gemini Live API, processes transcripts through Gemini for experience extraction, calls Gemini again for resume critiques and coaching context, writes structured data to Cloud Firestore, and serves it all through Firebase Hosting. On paper, that sounds like a recipe for a terrifying cloud bill.

I had set up pricing tiers early — partly because that's what you do with a SaaS product, and partly because I was genuinely nervous about subsidizing AI calls out of my own pocket. The plan was to charge enough to cover the infrastructure and maybe, eventually, turn a small profit.

So I did what any anxious founder would do: I pulled the actual numbers from the production GCP project.

The numbers told a different story

I ran a full cost analysis. Cloud Functions invocations, Firestore reads and writes, Firebase Auth sessions, Storage, Hosting bandwidth, Gemini API token usage — everything, pulled directly from Cloud Monitoring and the billing console.

The result surprised me.

Total estimated monthly cost for the entire platform was far lower than I expected. Not per user. Total.

Almost every service was comfortably inside Google Cloud's free tier. Firestore was using less than 1% of its daily free allowance — 360 reads per day against a 50,000/day limit. Cloud Functions had processed 722 invocations in a month against a 2-million monthly free tier. Firebase Hosting and Auth weren't even registering on the billing radar.

The only meaningful cost driver was the Gemini API itself — the LLM calls for coaching sessions, resume critiques, and experience extraction. Even there, the projected per-user cost at moderate usage stayed low enough that it didn't justify putting the core features behind a paywall.

I had been worried about going bankrupt. The reality was that Google Cloud's generous free tiers and Firebase's Spark plan had made the infrastructure much cheaper to run than I'd assumed.

The shift

That realization changed everything about how I thought about the product.

If the cost of serving a user wasn't the burden I had feared, why was I putting a paywall between people and a tool that helps them reflect on their careers? These are job seekers. Students. People between roles. People who are already stressed about money. Charging people to use an AI career coach started to feel wrong.

Not wrong in a business sense. The economics still worked. Wrong in a karma sense.

Free for individuals. Period.

So here's the pivot: My Job Odyssey is going free for individuals. All of it. The voice check-ins with Gemini Live, the career memory graph, the resume editor with live function calling, the resume tuner, the interview practice — no paywall.

I'm keeping rate limits in place — not to create artificial scarcity, but to prevent abuse and keep the platform sustainable. But there's no premium tier you need to unlock. If the tool can help you capture a career win, prepare for an interview, or update your resume, it should just do that. No credit card required.

So where does the money come from?

The real value My Job Odyssey creates isn't the resume — it's the graph. Every user who checks in weekly is building a rich, structured representation of their professional experience. Not a polished resume summary — a verified, STAR-format career graph extracted from real conversations about real work.

That's the future monetization: the B2B side. Enterprise coaching for companies. Recruiter access to the experience graph. Revenue from the organizations who want to find or develop talent — not from the individuals building their career story.

The individual users aren't the product. They're the point.

What's next

The next few weeks are about making the product better and spawning new initiatives. Building My Job Odyssey on Google Cloud taught me an enormous amount about deploying Gemini Live agents, managing real-time audio pipelines, and structuring a full Firebase stack for production. There are other problems that deserve the same treatment.

The hackathon gave me the impulse to start. The cost analysis gave me the confidence to keep going. And the decision to make it free gave me the clarity to focus on what actually matters: helping people remember their careers and tell their stories.

---

My Job Odyssey is free at myjobodyssey.com. No paywall. No trial period. Just a voice AI — powered by Gemini Live on Google Cloud — that helps you remember what you're worth.