Eric Smith Hits $10K MRR
Eric Smith announced that ProjectX has reached $10K in monthly recurring revenue (MRR). He shared a detailed expense breakdown and reflected on the implications of using AI tools to build the project, highlighting how non-coders can now create software MVPs more easily.
March 30, 2026$100 MRR Milestone
Original Announcement
š Hitting $10K rev a month now on ProjectX. Almost entirely "vibe coded".
Since people seemed so interested last time, here's a better expense breakdown of ProjectX:
āāāāāāāāāāāāāāāāāāāāāā
š Gross Revenue: $10,362
āāāāāāāāāāāāāāāāāāāāāā
š Expenses: $7,907.5
āāāāāāāāāāāāāāāāāāāāāā
* AI Video Models (Replicate) - $4,182
* Facebook Ads - $3,350
* Stripe Fees + Refunds - $932
* Servers + CDN - $219
* Gemini - $65
* OpenAI - $36
* Fal - $32
* Emails - $19
* Elevenlabs - $41
āāāāāāāāāāāāāāāāāāāāāā
š Profit: $1,480
I have a software engineering background working in FAANG, but I built 90% of this project using Codex+Claude, just to understand them better.
Clearly, these models are powerful enough to create profitable apps given the right direction.
However, it's a weird feeling not understanding the codebase that well, and the more you use it, the more you rely on it to add new features simply because you don't know your own code anymore.
Very interesting because now non-coders who are experts in their niche/domain can create useful software, at least at an MVP level. Used to be a huge barrier to entry that I think will come down.
Since people seemed so interested last time, here's a better expense breakdown of ProjectX:
āāāāāāāāāāāāāāāāāāāāāā
š Gross Revenue: $10,362
āāāāāāāāāāāāāāāāāāāāāā
š Expenses: $7,907.5
āāāāāāāāāāāāāāāāāāāāāā
* AI Video Models (Replicate) - $4,182
* Facebook Ads - $3,350
* Stripe Fees + Refunds - $932
* Servers + CDN - $219
* Gemini - $65
* OpenAI - $36
* Fal - $32
* Emails - $19
* Elevenlabs - $41
āāāāāāāāāāāāāāāāāāāāāā
š Profit: $1,480
I have a software engineering background working in FAANG, but I built 90% of this project using Codex+Claude, just to understand them better.
Clearly, these models are powerful enough to create profitable apps given the right direction.
However, it's a weird feeling not understanding the codebase that well, and the more you use it, the more you rely on it to add new features simply because you don't know your own code anymore.
Very interesting because now non-coders who are experts in their niche/domain can create useful software, at least at an MVP level. Used to be a huge barrier to entry that I think will come down.
Founder
Eric Smith
@ericsmith1302