The Real Cost of Building an AI Text-to-Speech Website: My Journey and SEO Lessons (2025)

👋 Introduction

Hi everyone! I’m a backend developer (mostly .NET Core + SQL Server).

This article has two main goals:

1. To share the real cost of building and running a live website.
2. To ask for insights and advice on how to do SEO effectively on Google.

🚀 From a Simple Tool to a Real Website

Earlier this year, I wrote a small Text-to-Speech (TTS) tool for a freelance project.

After trying it out, I realized this niche had great potential — not only for content creators but also for monetization through ads.

So, I decided to turn it into a real AI-based website.

I started coding in February 2025 and deployed the first version in May 2025.

💰 Actual Costs I’ve Spent

💡 Note:

When registering Google Cloud, use a low-balance card (30 K–40 K VND ≈ $1–2 USD) to avoid unexpected charges after the trial — by default, Google can charge up to 2 M VND (~ $80 USD) if billing isn’t disabled 😅.

📆 Development Timeline

May 18 → Jul 29, 2025:

1. The site had only one page — a basic TTS demo. Then I got lazy and left it 😅
2. After the Google Cloud trial expired, I shut it down.

Sep 26, 2025:

1. I decided to revive the project with the goal of learning and generating passive income.
2. Bought a Linux server, mail service, and redeployed everything.

Oct–Nov 2025:

1. Continued building new features:
2. 🧠 Speech to Text (renamed Chat AI) – integrated with OpenAI API.
3. 🖼️ Image-to-Text (OCR) – implemented with Python and EasyOCR.
4. 📰 Blog SEO – officially launched on Oct 14, 2025 (still low traffic 😅).

📈 Some Real Stats

From late October to early November 2025, the website stayed stable:

1. Around 50–100 users per day (GA4).
2. TTS generation requests ranged between 60–150 per day.
3. Clicks and impressions on Google Search Console are slowly increasing (see chart).

🧩 Lessons Learned After 6 Months

1. Don’t wait for perfection before launching.
2. I left the site idle for months when I could’ve collected early feedback.
3. Google needs time to understand your site.
4. It took about 2–3 weeks after adding sitemap, robots.txt, and JSON-LD for indexing to stabilize.
5. A blog really matters.
6. Each post is a new entry point for Google to crawl and increase domain trust.
7. SEO is a long-term game.
8. Traffic fluctuations are normal — what matters is consistency and real value.

🙏 What I’m Asking the Community

Right now I’m exploring:

1. How to SEO faster for tool-based utility websites.
2. Whether I should build a bilingual site (EN-VI) or focus on one language first.
3. Besides GA4 + Search Console, what tools are good for tracking user behavior?

If you’re experienced in web SEO or have done similar projects,

I’d really appreciate your insights and suggestions ❤️

💬 Conclusion

This post is both a personal milestone and a practical guide for anyone wanting to start a small AI web project.

Hopefully, it gives you a realistic picture of the costs (~ $170 USD total per year), the learning curve, and the patience needed to grow it.

Thanks for reading — feel free to share your thoughts or advice 🙌