2025 was the year ModernGrindTech stopped being a freelance side hustle and started being an actual studio. Twelve shipped projects, two pivots, four client engagements that converted to recurring work, one full marketing site rewrite (this one), and roughly $52k in invoiced revenue against $4.1k in operating costs. This is the quantified retrospective: what shipped, what flopped, what the numbers actually looked like, and what changes for 2026.
The Numbers
- Projects shipped: 12 (8 client builds + 4 internal tools)
- Gross revenue: $52,000 (rounded, recognized on delivery)
- Operating costs: $4,100 (infra, AI tooling, SaaS fees)
- Effective hourly rate: $94 across all engagements (calculated from invoiced hours)
- Median ship time: 18 days from contract signature to production deploy
- Slowest project: 11 weeks (third-party API blockers, not internal velocity)
- Fastest project: 2 days (eXp Black Lion, recruiting site, scoped for 4 weeks)
- Recurring engagements: 4 of 8 client builds converted to maintenance + new-feature retainers
- Refunds issued: 0
- Abandoned projects: 0
What Shipped
Client builds
Eight client engagements ranging from $499 to $4,999. The pattern that emerged: the smallest projects (under $1k) were the highest-stress because the scope expectations did not match the timeline, and the largest projects ($3k-$5k) were the most enjoyable because there was room to make real architectural decisions instead of cramming features into a weekend.
- Regal Title: Virginia title company website with RESPA compliance documentation, closing paperwork upload portals, and rate calculators. Scoped for 3 months by the agency they initially spoke with. I shipped in 3 weeks.
- 2K Service Plug: NBA 2K player marketplace built on Discord. Hit $10k in transactions its first quarter and converted to a recurring engagement.
- eXp Richmond: Recruiting platform for an eXp brokerage with branded training modules and lead capture. Scoped for 8-10 weeks by the previous bidder, shipped in 2.
- FlexTech (InflexTech): Premium photonics company website. Three.js R3F lab module visualizations, dark-mode-first design, cyan-purple gradient brand language.
- Holy Services: 2K marketplace with Stripe payments and Discord-bot fulfillment.
- Crown Money: Financial coaching platform with content-delivery and progress tracking.
- eXp Black Lion: Second eXp brokerage site, shipped in 2 days off the Richmond template.
- GTP World: Discord community marketplace with role-based access and event scheduling.
Internal tools
Four tools built to support the studio operation, each one paying for itself in time saved within the first month of use.
- Mission Control: Cron registry and execution platform. Replaced a tangled mess of system crontabs across two machines with a single web dashboard and audit log.
- X Engine: Twitter/X engagement automation that runs the @ModernGrindTech account through a daily cycle of replies, posts, and DM management.
- ClawFactory (MGT Factory): Content automation pipeline. Trends → hunt → clip → caption → safety → post across YouTube Shorts, TikTok, Instagram Reels, and X.
- MGT Suite: This site, the backend that powers it, and the admin dashboard. Rewritten from scratch in October 2025 and continuously shipped against since.
The Two Pivots
Pivot 1: From hourly to fixed pricing
The first six engagements were billed hourly at $95-$125/hr. The pattern was the same every time: client signs the contract, I estimate hours, I beat the estimate by 30 to 50 percent because of AI-assisted development, I bill the lower number and feel like I am leaving money on the table because the value delivered was the project, not the hours. The seventh engagement converted to fixed pricing at the median of what an agency would charge, with the explicit framing that the project ships in weeks not months. Same value, simpler conversation, no scope-creep negotiations. Every engagement since has been fixed price. The effective hourly rate jumped from $95 to $140+ on the same scope.
Pivot 2: From "developer for hire" to "shipping studio"
Through Q3 the inbound was framed as freelance work. Clients arrived with a feature request, I quoted hours, we negotiated. Q4 the framing changed to shipping studio: I quote a fixed price for a complete deliverable, not a feature list. The marketing site was rewritten around shipped projects rather than service categories. Inbound flipped from "can you do X by Friday" to "we have a project, what is your next slot." The same inbound channels, different conversion behavior because the framing changed.
What Did Not Work
Two product experiments shipped and immediately stalled. LaunchWise (an AI-assisted product-launch toolkit) hit a chicken-and-egg between needing real launches to validate the workflow and needing the workflow to attract real launches. MGT DevKit (a developer onboarding template) was technically clean but had no acquisition channel. Both stayed live, both still get the occasional click, neither generated revenue in 2025.
The lesson from both: technical execution is not the bottleneck. Distribution is. Internal tools paid for themselves because the user (me) was already there. Client builds paid for themselves because the user (the client) had a problem and a budget. Speculative products had neither.
The Tooling Story
2025 was when AI-assisted development stopped being an accelerator and started being the default. The numbers from the books:
- Claude Code: $200/mo for the full year, $2,400 total. Effective code-write rate when paired with the Claude Max model: roughly 4-6x baseline. Higher on greenfield, lower on legacy refactors.
- Cursor: $20/mo for the IDE integration, used as the editor of record across all projects.
- OpenAI/Anthropic API direct: ~$80/mo for production AI features inside client apps (chatbots, summarization, embeddings).
- Vercel + Neon: ~$80/mo combined across all hosted projects. The Vercel hobby tier carried the smaller projects; pro tier covered moderngrindtech.com and the X engine.
The interesting line is that AI tooling cost ($300-ish per month) is materially less than the time it saved per week. The math: if the tooling reduces 15 hours of routine work per week and the effective hourly rate is $140, that is $2,100 of reclaimed value per week against $300 of monthly cost. The return is obvious enough that the only question is whether to keep the savings as margin or reinvest into rate reduction. I chose margin for 2025; 2026 will mix both.
What Changes in 2026
Three things tightening up in 2026 based on what 2025 taught me.
First, no more sub-$1k engagements. They consume the same admin overhead as a $5k project (kickoff call, requirements doc, contract, invoicing, support handoff) without the margin to justify the time. The new floor is $1,499.
Second, retainer slots get cap. Recurring work was the single biggest revenue contributor in 2025 (4 of 8 builds converted), but it also crowded out new project capacity in Q4. The new structure is two retainer slots open at any time, with a waiting list when both are filled.
Third, public transparency expands. The 2025 retrospective is one piece. The /financials page (monthly waterfall) is another. The /pricing-honesty page (per-tier cost breakdown) is the third. The framing: agencies hide their math, MGT publishes it. That is the differentiation, and it gets reinforced quarterly with new public artifacts.
The 2026 Targets
Public targets for 2026 so they are easy to score against this time next year.
- Revenue: $120k gross. Roughly 2.3x 2025.
- Projects shipped: 16 (12 client builds + 4 internal tools). Slightly lower volume on the client side, larger average ticket size.
- Recurring revenue: 40% of total. Scaling the retainer side rather than the one-off side.
- New transparency artifacts: One per quarter. Q1 was /pricing-honesty, Q2 is this retrospective + /financials, Q3 is the public delivery dashboard.
- One sub-product launched and validated. Either MGT Suite as a paid platform or a new bet entirely. The criterion is paying users, not just tech ship.
If you are reading this in early 2027 and want to compare what landed against this list, every section of this site has the receipts: /case-studies for the shipped work, /financials for the monthly numbers, /changelog for the build log. The numbers should speak for themselves by then.