The $5K MVP Playbook: Stack, Scope, What We Cut

Here's what fits, what doesn't, the stack we don't negotiate at this price, and the kind of founder this tier is wrong for.
(Context for the price sheet: MVP cost in 2026.)
What you get for $5K
- One user persona: founder, customer, admin, pick one
- One core flow done well, end to end
- Auth (email + password or magic link)
- One paid plan via Stripe if money's involved
- Production deploy at your custom domain with monitoring
- Source in your GitHub from commit one
- Infra in your cloud account (Vercel or AWS) with environment separation
- One round of post-launch scope adjustments (1 to 2 days) inside the price
Time to first deploy: 7 to 10 calendar days.
What you don't get
- Multi-tenant data isolation (workspaces, teams): that's $20K+
- Admin panel beyond the bare minimum to run the thing
- Multi-step Stripe billing (usage metering, seat-based, dunning): basic checkout only
- SSO (SAML, OIDC)
- Mobile app: web only
- AI features beyond a single LLM call: no RAG, no agents, no evals
- Real-time collaboration
- Internationalization
- Design system from scratch: we compose with shadcn/ui
- Custom illustrations or branded visuals
- Marketing site: we build the product, not the landing page
- A dedicated project manager: you talk to the engineer directly
If you need any of those at v1, $5K is the wrong tier. We'll say so on the call.
The stack we don't negotiate at this price
At $5K we don't bill for stack-selection meetings. The stack is fixed:
- Next.js (App Router, RSC, current LTS): frontend, API routes, deploy story in one piece
- TypeScript strict mode end to end
- PostgreSQL via Drizzle or Prisma
- Vercel for hosting (or your AWS account if you ask)
- Stripe for payments
- Resend for transactional email
- Tailwind + shadcn/ui for design
- Sentry for error monitoring
- Vercel Analytics + GA4 for product analytics
We'll deviate for free if you have an existing stack we should match: Supabase, Clerk, Postmark, Neon. We'll deviate for cost-credible reasons. We won't deviate to chase a new framework or to satisfy a Twitter argument.
This is the same stack documented in How we use Claude Code in production. The economics depend on the stack being known.
How $5K is even possible
Modern AI tooling collapses the boilerplate. A senior engineer with Claude Code spends maybe 30 minutes setting up auth (was 4 hours pre-2024), 1 hour on Stripe wiring with webhooks (was 2 days), 2 hours on the admin scaffold (was 1 day). Architecture, schema, security model, edge cases still take the same human-bound time. The total compresses to about 30 to 40 hours of engineering for a one-flow MVP.
At a $130/hour effective senior rate, that's $4K to $5K of work. We round to $5K to keep the math clean and leave room for one post-launch scope round.
Stress-test your idea against the playbook
Ask ChatGPT to map your scope onto this playbook. It'll surface which pieces fit at $5K, which push you into a higher tier, and where you should be having different conversations.
“|”
Five signs this tier is wrong for you
You don't have a tight spec. A $5K engagement assumes a one-page spec, not a 20-page PRD that hasn't been written. If you're still arguing whether the core flow should be onboarding-first or value-first, talk to users before writing code.
You're past validation. $5K is a validate-the-hypothesis price. If you have 200 paying users on a no-code prototype and need the real thing, you want $20K to $30K: multi-tenant, billing edge cases, real admin.
You need a custom design system. We compose with shadcn/ui. If the product needs a unique visual identity at launch, the design budget is separate and material.
The product is regulated. HIPAA, PCI L1, SOC 2 from day one. Compliance is a labor floor that doesn't compress with AI. Different conversation.
You want to chat through scope weekly. Flat price assumes stable scope. Discovery-style engagements should be hourly or at a higher tier.
Have a one-page spec ready?
Send your email and we'll set up a 30-minute call within 24 hours. After the call you get a written quote, and an honest read on which tier actually fits, even when it's bigger than $5K.
How to make $5K work
If this is the right tier:
- One-page spec. Bullet points are fine. We'll send back questions; answer those and the quote follows.
- Thirty minutes a day during the build window. Most days no question. On the day there is one, we need a quick answer to avoid blocking.
- No scope additions mid-build. One post-launch round is included. Mid-build additions push the price into the next tier.
- Honest post-launch call. After day 14 you'll have a working product. Decide: keep us on retainer to iterate, hire your first engineer to take over, or freeze and run with what shipped. All three are fine.
Per-city pages
Engagement model detail: Why we don't take equity.
Enjoyed this article? Share it with others
Related Posts

Getting Your Code Back: What to Demand Before You Walk Away
Most engagement contracts assign IP to the client. Most agencies still gatekeep the code on the way out. Here are the seven things to demand before you sign off the final invoice, and the order to demand them in.

Burning Tokens Is Not Shipping
A company gave every engineer unlimited Claude Code and put up a leaderboard for who burned the most tokens. The post is at 1,000 upvotes. The top comments are about gaming the leaderboard. The leaderboard is the product.

MVP Cost in 2026: Real Numbers From $5K to $50K
Most agency quotes are still pretending it's 2023. Here's what an MVP actually costs in 2026: five tiers, what each one ships, where the price comes from, and the work we won't take at the bottom of the range.