SaaS boilerplate vs building from scratch: costs, timelines, and how to decide
Should you use a SaaS boilerplate or build from scratch? Compare real timelines, hidden costs, when DIY wins, and how TurboStarter gets you to billing faster without rebuilding auth.
If you are deciding between a SaaS boilerplate and building from scratch, you are not choosing between "lazy" and "serious." You are choosing where to spend your first 8–16 weeks: on infrastructure every SaaS needs, or on the product only you can build.
Short answer
For most founders, a production SaaS boilerplate is the better default. Building auth, billing webhooks, teams, admin tooling, and marketing from tutorials typically takes 10–16 weeks of focused work before you ship differentiated features. A paid boilerplate like TurboStarter collapses that to days or a few weeks, because you get full source code, a documented monorepo, and the boring 80% already wired — web, mobile, and browser extension clients included. Build from scratch only when you have unusual compliance needs, a research mandate, or infrastructure that is itself your competitive advantage.
What is a SaaS boilerplate?
A SaaS boilerplate (also called a SaaS starter kit) is a production-oriented codebase that ships the shared foundation of subscription software: authentication, billing, database schema, API layer, admin UI, emails, marketing pages, and deployment patterns. You buy or clone it once, own the source, and extend it with your product logic.
It is not a no-code template. It is not a locked hosted platform. Modern kits like TurboStarter are Turborepo monorepos with strict TypeScript, shared packages (auth, billing, db, api), and docs so your team — or AI coding agents — can navigate the repo without reinventing folder structure.
What "building from scratch" actually means
"From scratch" rarely means writing a TCP stack. In practice it means:
- Picking a framework (usually Next.js)
- Wiring auth yourself (sessions, OAuth, MFA, password reset)
- Integrating Stripe, Lemon Squeezy, or Polar with working webhooks
- Designing multi-tenant orgs, roles, and invitations if you sell B2B
- Building admin tools, emails, landing page, blog, legal pages, SEO
- Adding tests, monitoring, and deployment before you charge money
That is months of work that does not differentiate your product. Customers pay for your workflow, data model, or AI — not for another login form.
The real comparison is not boilerplate vs blank folder. It is boilerplate vs free starter + months of glue code vs full DIY.
Time comparison: realistic timelines (not marketing "Day 1")
Read the timelines carefully
Competitor pages love tables where every row says "Day 1" with a boilerplate. That is misleading. Even with a good kit, you still configure env vars, branding, pricing, and your core feature. The win is measured in weeks removed, not hours fantasized away.
Below are estimated timelines for a solo developer or two-person team shipping a billable B2B SaaS MVP. Your velocity varies; treat these as planning ranges, not guarantees.
| Workstream | With TurboStarter (or similar paid kit) | Build from scratch | Free starter only (Next.js SaaS Starter) |
|---|---|---|---|
| Repo + monorepo setup | 0–1 day (included) | 3–7 days | 1–2 days |
| Auth (email, OAuth, MFA, passkeys, magic links) | 1–3 days (configure + customize) | 3–6 weeks | 1–2 weeks (partial) |
| Organizations, invites, RBAC | 2–5 days | 2–4 weeks | Often not included |
| Billing + webhooks + entitlements | 3–7 days | 3–5 weeks | 1–3 weeks |
| API layer + typed clients | 1–3 days | 2–4 weeks | Minimal |
| Admin dashboard + impersonation | 1–3 days | 1–3 weeks | Rarely included |
| Emails (transactional + templates) | 1–2 days | 1–2 weeks | Basic |
| Marketing (landing, pricing, blog, SEO, i18n) | 2–5 days | 2–4 weeks | Partial |
| Mobile app (Expo) or browser extension | 1–2 weeks (shared API) | 6–12+ weeks each | Not included |
| Storage, analytics, monitoring hooks | 1–3 days | 1–2 weeks | DIY |
| Total to billable MVP | ~1–3 weeks | ~10–16 weeks | ~4–8 weeks |
The gap is not "install vs invent." It is production edge cases: failed payments, seat limits, session expiry, invite flows, webhook idempotency, and plan-gated API routes. Boilerplates encode those lessons so you do not pay tuition on launch week.
For the full feature map behind these rows, see our SaaS stack guide.
Cost comparison: license vs opportunity cost
Boilerplate approach (paid kit)
| Item | Typical range |
|---|---|
| Boilerplate license (one-time) | ~$200–$500 for indie-focused kits; team tiers higher |
| TurboStarter Core Kit | $249 one-time (lifetime updates on core) |
| Hosting (Vercel, Railway, Fly, etc.) | $0–$50/month early stage |
| Database (Neon, Supabase, RDS) | $0–$25/month early stage |
| Email (Resend, etc.) | $0–$20/month at low volume |
| Domain | ~$10–15/year |
Estimated first-year cash cost: roughly $400–$1,200 plus your time building product features.
Build from scratch
| Item | Typical range |
|---|---|
| Developer time (10–16 weeks) | At $75/hr × 400–640 hrs → $30,000–$48,000 outsourced |
| Your own time (opportunity cost) | 2.5–4 months full-time you are not selling or validating |
| Same hosting, email, domain | ~$200–$800/year |
Estimated first-year cost: tens of thousands in cash or months of calendar time before differentiated value ships.
The breakeven math is blunt: if a $249 license saves even two weeks of founder time, it pays for itself unless your hour is worth less than a few dollars. Most funded startups still buy kits because speed to learning beats auth pride.
What you still build with a boilerplate (important)
A good boilerplate does not build your product. You still own:
- Core workflow and data model
- UX that matches your niche
- Integrations specific to your domain
- Go-to-market, support, and positioning
- Compliance paperwork (privacy policy content, DPAs, SOC2 path if needed)
What you stop rebuilding: session security, Stripe webhook handlers, org membership tables, pricing page wiring, and admin screens to debug a stuck customer.
TurboStarter documents exactly what is included — auth matrix, billing providers, organizations, Hono API, admin, marketing CMS, Docker local dev, and AI development rules for Cursor and Claude Code.
When a SaaS boilerplate is the right choice
Use a boilerplate when most of these are true:
- You need paying users soon — validation beats architectural purity
- You are solo or a small team — calendar time is your scarcest resource
- Your moat is not infrastructure — the value is workflow, data, or distribution
- You sell B2B or may add teams — orgs, RBAC, and admin are table stakes
- Your roadmap includes mobile or a browser extension — a monorepo saves a rewrite
- You use AI coding tools — structured repos with
AGENTS.mdand package boundaries outperform a tutorial mashup - You want PostgreSQL ownership — self-hosted auth and data without Auth0 lock-in early
Boilerplates are especially strong for indie hackers, agency client launches, and seed-stage teams that must show traction before hiring specialists for every layer.
When building from scratch makes sense
DIY is rational when:
- Compliance requires custom control — regulated health, finance, or government contexts where every dependency is audited line-by-line
- Your core product is developer infrastructure — you are selling auth, billing, or deployment itself
- You have a large platform team and a long runway — learning and standardization across dozens of services is the goal
- No boilerplate matches a hard constraint — exotic tenancy models, on-prem-only, or non-Postgres mandates you cannot adapt
- The journey is explicitly the product — education, portfolio, or research where rebuilding is the point
Even then, many teams start from a free starter (Next.js SaaS Starter) rather than a truly empty repo — that is still "not from scratch," just more DIY assembly.
Three paths compared
Paid boilerplate
Multi-platform SaaS, B2B, AI-ready products. Billable MVP in ~1–3 weeks. Tradeoff: upfront cost and learning the repo layout.
Free starter + glue
Learning, tight budget, web-only. ~4–8 weeks. Tradeoff: missing orgs, admin, mobile; webhook bugs.
Full scratch
Unique infra, compliance, large teams. ~10–16 weeks. Tradeoff: never shipping; reinventing known failures.
| Path | Best for | Time to billable MVP | Main risk |
|---|---|---|---|
| Paid boilerplate (TurboStarter) | Multi-platform SaaS, B2B, AI-ready products | ~1–3 weeks | Upfront cost; learn the repo layout |
| Free starter + glue | Learning, tight budget, web-only | ~4–8 weeks | Missing orgs, admin, mobile; webhook bugs |
| Full scratch | Unique infra, compliance, large teams | ~10–16 weeks | Never shipping; reinventing known failures |
For a kit-by-kit shootout, read our SaaS starter kit comparison. For stack depth, read the SaaS stack guide.
Common concerns about boilerplates
You are locked in only if you cannot read TypeScript. Paid kits ship full source — no runtime dependency on the vendor. TurboStarter uses standard tools (Next.js, Better Auth, Drizzle, Hono) documented across the ecosystem. If the vendor disappeared tomorrow, your app keeps running; you lose updates, not ownership.
Mitigation: prefer kits built on popular OSS, with clear package boundaries and migration docs.
Understanding grows faster when you read working patterns than when you debug a broken OAuth redirect at 2 a.m. Start from documented modules — auth, billing, API recipes — and change one vertical at a time.
If learning is the primary goal, budget time anyway; a boilerplate just moves learning from "implement MFA" to "extend MFA for my use case."
Scale is a function of database design, caching, and ops — not whether you wrote the login page. TurboStarter targets PostgreSQL, serverless-friendly APIs, and patterns used in production SaaS. Scale problems at 10k users look like queries and jobs, not boilerplate brand names.
Some kits are bloated. Evaluate package graph and delete paths: can you remove mobile, extension, or AI packages cleanly? TurboStarter's monorepo is modular — use web only today, add Expo when the roadmap demands it.
Investors fund traction and velocity. Using a starter to reach revenue or LOIs faster is a rational capital allocation. What hurts fundraising is missing metrics because you were rebuilding Stripe webhooks.
How to evaluate a SaaS boilerplate
Before you buy, score kits on what actually blocks launches:
- Auth depth — MFA, passkeys, OAuth, magic links, org plugins (Better Auth is a strong signal)
- Billing realism — multiple providers, customer portal patterns, webhook tests
- B2B surface — orgs, invites, roles, admin impersonation
- API story — typed routes shared across web, mobile, extension (Hono or tRPC)
- Platform scope — web-only vs monorepo with mobile/extension
- Docs + AI readiness —
AGENTS.md, skills, recipes for AI agents - Maintenance — changelog, Discord, issue response time
- Demo you can click — web demo, store listings if mobile matters
Skip kits that list features without docs pages for each. Marketing bullets without migration paths are a red flag.
Why founders pick TurboStarter over scratch (and over thinner kits)
TurboStarter is built for shipping a product system, not a login page:
- Multi-platform monorepo — Next.js web, Expo mobile, WXT extension sharing one Hono API
- Production auth — Better Auth with sessions, OAuth, MFA, passkeys, orgs (auth docs)
- Billing flexibility — Stripe, Lemon Squeezy, Polar with UI and webhooks (billing docs)
- Operator tooling — super-admin, user search, impersonation (admin docs)
- Growth surface — landing, blog, docs, SEO, i18n, emails (marketing)
- AI-assisted development — AI development guide, skills, and conventions for Cursor/Claude Code (Cursor best practices)
Compared to supastarter, TurboStarter emphasizes one cohesive monorepo with mobile and extension clients and Hono — not picking among parallel framework SKUs. Compared to ShipFast, it trades "smallest possible web MVP" for B2B depth and future platforms without a second purchase.
You are not buying hype. You are buying back 10+ weeks on solved infrastructure so sprint one targets your unique feature.
Decision checklist: which path should you take?
Answer yes/no:
| Question | Yes → lean boilerplate | No → consider more DIY |
|---|---|---|
| Do you need billing in the next 30 days? | ✓ | |
| Will you sell to teams or companies? | ✓ | |
| Is mobile or an extension on the 12-month roadmap? | ✓ | |
| Are you using AI agents to write code? | ✓ | |
| Is your differentiator unrelated to auth/payments? | ✓ | |
| Do auditors require you to author every security control? | ✓ | |
| Is building the framework the actual product? | ✓ |
If you checked mostly left: start with TurboStarter, run pnpm services:setup and pnpm dev, and ship your differentiator this month.
If you checked mostly right: plan a custom architecture — but still consider a free starter for individual layers instead of reinventing all of them.
Frequently asked questions
Yes, for most founders. A one-time license in the $200–$500 range is cheap relative to 10–16 weeks of infrastructure work. The kit pays for itself if it accelerates launch or first revenue by even a couple of weeks.
Yes. You receive full source code. Rename components, replace providers, delete apps you do not need. Treat the kit as a maintained starting commit, not a hosted sandbox.
The official Next.js SaaS Starter is an excellent minimal reference: auth, basic billing, dashboard. Paid kits add orgs, admin, emails, marketing, mobile/extension paths, monitoring hooks, and AI-friendly docs — the surface between "demo" and "business."
Many teams reach a billable MVP in 1–3 weeks: days for env, branding, and pricing; the rest on unique product logic. DIY from scratch is typically 2–4 months for the same infrastructure breadth.
No. SEO depends on your marketing pages and content — TurboStarter includes SEO patterns. Performance depends on how you query and render — the starter gives sane defaults; your feature code determines Core Web Vitals.
You keep the code. Pin dependencies, fork internally, and migrate packages on your schedule. Prefer vendors with public changelogs and active communities to reduce that risk.
Yes. Use the saved months for distribution, hiring, and domain features. Disclose stack choices to investors if asked; traction matters more than auth authorship.
Boilerplate, especially with documented monorepos and AI rules. Agents need boundaries. TurboStarter's AI development docs map where auth, billing, and features live — reducing hallucinated parallel architectures. Add the AI kit when the product needs LLM workflows.
The verdict
For most SaaS founders, building from scratch is the expensive default, not the virtuous one. Authentication, billing webhooks, teams, admin, and marketing are solved problems. A boilerplate buys calendar time and reduces launch risk.
Build your differentiation. Stop rebuilding the thousandth login form.
If you want a production monorepo with web, mobile, extension, auth, billing, orgs, and operator tooling already integrated, start with TurboStarter. Read the docs, compare TurboStarter vs supastarter, and spend your first real sprint on what customers will pay for.
Renovate vs Dependabot - what's the best tool to automate your dependency updates?
Compare Renovate and Dependabot for dependency automation. Learn key features, benefits, and strategies to choose the right tool for your monorepo.
Self-host your Next.js Turborepo app with Docker in 5 minutes
Learn how to containerize your Next.js Turborepo app with Docker, optimize the image, and deploy it to any environment in minutes with practical examples.
