Published on

Building Senimaya.com: ez to build, hard to sell

Authors
  • avatar
    Name
    Haikal Tahar
    Twitter

Introduction

i thought the hardest part of building senimaya.com would be the engineering. turns out, the hardest part is money. getting users to pay is a whole different game compared to getting the tech to work.

this post is about what i built, what i learned, and the challenge of turning a working product into a business.

Table of Contents

1. Engineering Phase

as an engineer, i was excited.

  • got image generation working by ready to use inference provider(fal.ai, openrouter.ai)
  • built a credit system to control cost, starting price $7
  • added an ai gallery to showcase results
  • only cost $10 for the domain but image gen start to get expensive the more image is produce

from a technical perspective, senimaya.com worked. i had a running product, fast outputs, and features people could use.

2. Product Phase

this is where reality hit.

  • i got some early visitors, but very few signups
  • users left comments like: “why should i use this instead of ChatGPT? GPT is free, yours i need to pay”
  • i realized “working” isn’t the same as “valuable”
  • most people didn’t care about infra speed or credit systems — they wanted clear reasons to pay

engineering gave me a product. product work made me face the market.

3. Key Learnings

  • free vs paid: competing with free tools (like ChatGPT) means you need super clear differentiation
  • attention ≠ adoption: comments and feedback don’t always convert to paying users
  • engineering joy vs market reality: what feels like a breakthrough in code can feel invisible to users
  • value story matters more than features
  • analytics matter: i learned to use tools like posthog.com to understand how users interact with the site
  • feedback loops: i set up a place for users to give feedback and suggest features at senimaya.canny.io

4. Challenges

as an engineer

  • easy to spend weeks improving infra no one notices
  • building features no one asked for because they were “fun”
  • shipping doesn’t equal revenue

as a product builder

  • convincing users why this exists when “ChatGPT is already free”
  • translating vague feedback (“make it better”) into actionable changes
  • the loneliness of no traction despite weeks of building

5. Future Improvements

  • double down on use cases GPT doesn’t solve well (product photography, fashion try-on, interior design)
  • show side-by-side comparisons: “here’s what GPT makes vs here’s what Senimaya makes”
  • add free tier or trial to reduce friction for first-time users
  • focus marketing on niches that need visuals, not general users

6. Conclusion

engineering senimaya.com was hard but doable. monetizing it is another beast. the gap between a product that works and a product people pay for is wide.

the biggest lesson so far: building is only half the journey. the real challenge is convincing strangers that what you built is worth their money.