I write this at 11:47pm with a half-cold flat white and a Stripe dashboard open in a side tab. That tab is the entire point. If you're an indie hacker or a solo founder in 2026, your job is brutally simple. Build a thing, charge for it, talk to humans, repeat. Your tooling should make all four feel like one motion, not four jobs.

The good news is that the stack has gotten silly cheap. The bad news is that the marketing pages all sound the same, every tool now has an "AI" in its sidebar, and you can easily burn $400 a month before you ship a single feature. So this is my actual stack: what I run, what I cancelled, and what I think you should pay for.

If a tool can't pay for itself in saved hours within thirty days, I cancel it. Indie hackers shouldn't be running enterprise software on a side-project budget.

What "indie-friendly" actually means in 2026

Before I list anything, let me be specific about the bar. A tool earns a slot in a solo founder's stack only if it does at least three of these well: usable free tier, no sales call, no minimum seat count, fair pricing under fifty users, and an API that doesn't require a partnership PDF.

That fourth point is where most of "AI" goes to die. Half the new tools in 2026 are wrappers with a $79/seat tag and zero leverage when you're alone. Avoid those. The ones that survived my last cleanup are below.

I group them by job rather than category. You don't wake up needing a "vector database." You wake up needing to ship a feature, get paid, or stop a churn problem. Tools should map to verbs, not categories. If you want a wider menu before settling, the running shortlist on the best AI tools index is where I keep notes between cleanups.

Build: writing the actual code

The single biggest unlock of the past 18 months wasn't a new framework. It was the shift from autocomplete to agentic coding. You stop typing functions. You describe outcomes. The tool reads your repo, edits ten files, runs the tests, and shows you a diff. Once you've felt it, going back is like switching from a dishwasher to handwashing.

My pair: Cursor in the IDE for tight, file-by-file work, and Claude Code in the terminal for big, sweeping changes ("rename this domain concept everywhere, update the migrations, regenerate the types"). They overlap. I don't care. The combined cost is less than my annual JetBrains license used to be.

$40/mo

my entire AI coding budget: Cursor Pro plus a Claude subscription with Code access

Cursor is twenty bucks a month for Pro and it pays back in the first hour of any non-trivial week. The tab-completion model alone has saved me from typing thousands of lines of boilerplate Svelte and Drizzle. If you're choosing between IDE-style assistants, the Cursor vs Windsurf comparison goes deeper on which feels right per use case. Windsurf is the cheaper alternative; Cursor still wins for me on raw context handling.

Claude Code is the one I'd defend most aggressively. It runs in your terminal, has full filesystem access in the project, and treats your codebase like a colleague's. I run it for migrations, refactors, test backfills, and the kind of yak-shaving you usually procrastinate. Once you've handed it a TODO list and watched it just do the list, you stop seeing tools the same way.

If you want a curated list of build-side options, including the ones I tried and dropped, the best developer tools shortlist mirrors the way I think about this segment.

v0 for the front-end first draft

For UI, I lean on v0 from Vercel. It generates clean Tailwind and shadcn components from a screenshot or a sentence. Is it production-grade out of the box? No. Does it cut a four-hour design pass to forty minutes? Yes. I treat its output like a junior designer's wireframe: useful 70% of the way, then I rewrite the rest.

Ship: getting the thing to a URL

The easiest way to lose a weekend in 2026 is to argue with yourself about hosting. Don't. If you're solo, the answer is one of two stacks.

For most apps, Vercel on the front, Supabase on the back, done. Push to main, your code goes live in 90 seconds, you have auth, postgres, storage, edge functions, and row-level security without ever opening AWS. You'll pay zero until you have actual users. When you cross the free tier you'll happily pay because it means you have users.

For "I want servers I can ssh into" people, a cheap VPS plus Coolify is fine. But if you're new at this, please don't. The hours you'll spend debugging a TLS cert at 2am are hours you could have spent on the actual product. I learned this the hard way more than once.

Email matters more than people admit. Use Resend. It costs essentially nothing for the volumes a solo founder hits, the React Email templates feel native, and the deliverability has been the best I've used since the old Postmark days.

Hosting is a solved problem in 2026. If you're still arguing about it on day one, you're hiding from product work. Pick Vercel and Supabase, ship, move on.

Get paid: Stripe, or one of the merchants of record

Three real options. Stripe for control and the lowest fees. Lemon Squeezy if you want a merchant of record so you don't have to deal with EU VAT, sales tax in 47 US states, and Australian GST. Polar if you're shipping mostly to developers and want a clean, modern alternative that handles the same MoR headaches.

I run Stripe for the SaaS and Lemon Squeezy for the digital goods. Yes, it's two integrations. Yes, the tax-handling savings are worth it. Anyone who tells you tax is "easy, just use Stripe Tax" hasn't actually filed in three jurisdictions yet.

BEFORE

9 hrs

monthly tax and invoicing busywork

AFTER

35 min

with a merchant of record handling it

The Stripe AI features added in 2025 (agentic checkout, invoice intelligence, fraud reasoning) are quietly excellent but mostly invisible if you set up the basics well. That's the right kind of AI. The kind that just makes the existing thing more correct.

Talk to users: support, scheduling, and ops

Solo founders don't have a CS team. You are the CS team. The trick is to make every conversation either fast or unnecessary.

For 1:1 calls and demos, Cal.com. Self-hostable if you want, hosted if you don't, free tier covers a single founder forever. The booking page customization in 2026 is genuinely good, and the API is generous if you want to embed it inside your app. Stripe-style polish, open-source heart.

For tracking issues, requests, and the public-facing roadmap, Linear for internal and Featurebase for the public side. Linear's $10/seat is the rare price tag I never resent. It's the only PM tool I've used where I open it because I want to, not because I have to.

For docs and writing, Notion still wins for me, even with the field full of upstarts. The AI features got actually useful in late 2025 (summarizing meeting notes, auto-tagging tasks, drafting outlines). Just keep them turned off until you've decided what your wiki actually is, otherwise the LLM will helpfully turn your knowledge base into mush.

If you want to see how this generalizes beyond me, the curated list of tools for solopreneurs includes a few that I rotated out but that other founders swear by, particularly for content-heavy products.

Grow: analytics, email, audience

This is the section where every "best of" list gets boring, so I'll be blunt.

For product analytics, PostHog. It's the only tool in this category that grows with you from "I just want to know if anyone's using feature X" to "we have funnels, session replay, feature flags, and A/B tests." The free tier is real. Self-host if you're paranoid; otherwise the cloud version is fine and includes the AI assistant for querying events in plain English, which I use more than I'd admit.

For lightweight, GDPR-friendly site analytics, Plausible or Fathom. They're functionally equivalent. Pick whichever owner you like more on Twitter; both are bootstrapped, both are good. I run Plausible.

For audience and email, ConvertKit if you're a creator/newsletter person, Beehiiv if you want growth tooling baked in (referrals, recommendations, ads). I'd start on Beehiiv in 2026. The Boost network alone is worth it for a new newsletter, and the AI subject-line and segmentation tools are not gimmicks.

If you're under 1,000 newsletter subscribers and reaching for a $99/mo email tool, you've made a mistake. Beehiiv free tier or ConvertKit's free plan will carry you through the first thousand readers. Spend the saved $99 on Cursor and ads.

One last one in this bucket: ChatGPT still has a place even if you live inside Claude all day. The voice mode and the deeper image tools are genuinely better for some research tasks, customer-call-prep, and casual thinking partner duties. I keep a $20 sub on it specifically for that. The general-purpose AI category gets messier every month. If you want a clean comparison, the best AI tools page tries to keep it honest.

The full stack, ranked by what I'd cut last

If a hurricane hit and I had to rebuild my stack on a $50/month budget, here's the order I'd add things back in.

Cursor or Windsurf, because the leverage is too high to skip.
Stripe or Lemon Squeezy, because you can't ship a SaaS without payments.
Vercel and Supabase free tiers, so hosting stays at $0 until traction.
Resend for transactional email, $0 to start.
PostHog free, because you cannot fly blind on a paid product.
Cal.com hosted free, for booking calls without back-and-forth emails.
Plausible or Fathom, once you have a marketing site worth measuring.
Claude Code, which pays for itself the first time you run a 30-file refactor.
Linear, when your "to do" list outgrows a Notion page.
Beehiiv or ConvertKit, when you actually have something to send.
v0, when UI quality starts to matter more than UI speed.
Featurebase, when users start asking for the same feature three times.

That ordering is not a hierarchy of "good." It's a hierarchy of what a brand-new solo founder would feel the absence of first. You'd notice no Cursor by lunch. You wouldn't notice no Featurebase for two months.

Honest tradeoffs and small admissions

Stripe vs Lemon Squeezy vs Polar is not a clean call. If you ship to other developers and you're confident in handling tax stuff yourself, Stripe wins on every metric except tax. If you're not, the ~5% premium on Lemon Squeezy or Polar is the cheapest accountant you'll ever hire.

Cursor vs Windsurf is genuinely close in 2026. Cursor still has the edge on context window and large refactors, in my hands. Windsurf is meaningfully cheaper. If you're cost-sensitive, start there. The side-by-side breakdown is the version I'd send a friend.

PostHog vs Plausible isn't really a vs. PostHog is product analytics (events, funnels, sessions). Plausible is site analytics (pageviews, referrers, lightweight). I run both, on different surfaces, and so should you when traffic justifies it.

If you want to keep things even leaner, the free developer tools shortlist is a good companion to this post. It's where I track the tools that have a real, indefinite free tier rather than a 14-day tease.

Who shouldn't use this stack

If you're building enterprise B2B with a six-month sales cycle, ignore most of this. You need things like SOC 2 dashboards, scoped RBAC, and SSO that "just works", and that's a different shopping list. The SaaS-specific tooling list covers more of that side, including a few paid options I wouldn't recommend to a solopreneur but that earn their keep at scale.

If you're regulated (healthcare, finance, government), you'll need vendors with the right certifications and BAAs and that's a much narrower field. Most of the tools above can get there, but you should not rely on a generic blog post (this one) to tell you which.

If you're a hobbyist who wants to build for the love of it and never charge: skip the payments and analytics tier, keep build and ship, and you'll be fine on $0.

What changed for me in the last six months

Three concrete shifts. I cancelled Notion AI and use Claude in a side window instead, which feels more flexible. I replaced a custom Resend and Postgres tagging hack with a small ConvertKit and Beehiiv combo for two different audiences. And I moved one project's billing from raw Stripe to Polar specifically to outsource the EU tax hellscape, which saved me a quarterly headache for a reasonable percentage.

Stack hygiene is its own discipline. Every quarter I open the credit-card statement, look at every recurring charge, and ask the same question: "if this was twice the price tomorrow, would I still pay?" Anything that gets a "hmm" gets cancelled. It's how I keep the stack honest.

Closing: pick a stack, ship a thing

The boring conclusion of every "best of" list is "it depends." But honestly: it doesn't, that much, anymore. In 2026, the indie stack has converged. Cursor or Windsurf for code. Vercel and Supabase for hosting. Stripe or a merchant of record for money. Resend for email. PostHog for product. Plausible for site. Cal.com for calls. Beehiiv for audience. Done.

The differentiation is not in the tools. It's in what you build with them, who you build it for, and whether you actually charge people for it. Pick the stack, stop researching, and go ship the thing. If you don't, someone using a worse stack will, and you'll find that very irritating.

If you want a starting list to compare against without leaving the site, the curated AI tools list, the developer tools list, and the solopreneur tools list are the three I update most often. Start there, then close the tab and go open your editor.

Insights & Resources

Best AI tools for indie hackers and solo founders in 2026

Stay Updated