Indie startup tech stack: what 1,354 live sites run in 2026
What tech stack do indie startups use? We checked 1,354 live sites on 10 November 2026. Next.js, Vercel, Supabase, Clerk and Stripe lead wherever we could see them.
Lourdes Paul Agilan
On this page
- Key findings
- What tech stack do indie startups use?
- Which framework do indie startups use?
- Where do indie startups host?
- Supabase or Firebase? What indie startups use for databases
- Auth and payments: Clerk and Stripe
- Do sites that make revenue use a different stack?
- What founders say they use
- What GitHub and developer surveys add
- What this means if you’re picking a stack
- How we measured it
- Caveats
- FAQ
- What tech stack do indie startups use?
- Do most indie startups use Next.js?
- Is Supabase more popular than Firebase for indie startups?
- Where do indie startups host their sites?
- Do indie startups use Stripe?
- Cite this research
- Sources
Short answer: among the sites where we could see a layer, the most common indie startup tech stack is Next.js on Vercel, with Supabase for the database, Clerk for sign-in and Stripe for payments. Next.js ran 71.1% of detected frameworks, Vercel hosted 52.6% of sites with a named host, Supabase held 71.3% of detected databases, Clerk 70.8% of detected auth, and Stripe 63.1% of detected payments.
Those are shares of the sites where we could see each layer, and coverage varies: 57.5% of sites for frameworks, 3.5% for auth. So the numbers don’t mean 71% of all indie sites use Next.js. They show which tool wins when you can see it.
We opened 1,354 live indie sites in a real browser on 10 November 2026 and read what each one loads. We also looked at founders’ own stack lists, GitHub dependency files and three public surveys, and report them separately.
Disclosure: We built Indie Torch with Astro on Cloudflare, so we use one of the minority stacks below. Every figure comes from the public sources listed at the end or from our own browser check of public sites.
Key findings
- Next.js ran 71.1% of detected frameworks (553 of 778 sites). Astro came second at 6.9%.
- Vercel hosted 52.6% of sites with a named host (438 of 833).
- Supabase held 71.3% of detected databases (77 of 108). Firebase and Firestore had 25.0%.
- Clerk held 70.8% of detected auth (34 of 48), and Stripe held 63.1% of detected payments (53 of 84).
- Next.js on Vercel was the top pairing: 326 of 537 sites (60.7%) where we could see both.
- Sites that make verified revenue lean harder on that stack: Next.js ran 82.7% of detected frameworks there, against 66.4% on launch sites.
- We could see a framework, a database and a host together on only 50 of 1,354 sites (3.7%).
What tech stack do indie startups use?
Here’s each layer, how many sites showed it, and who leads. “Share of detected” means the share among sites where we could see that layer.
| Layer | Sites where we could see it | Most common | Share of detected |
|---|---|---|---|
| Framework | 778 of 1,354 (57.5%) | Next.js | 71.1% (553 of 778) |
| Named host | 833 of 1,354 (61.5%) | Vercel | 52.6% (438 of 833) |
| Database | 108 of 1,354 (8.0%) | Supabase | 71.3% (77 of 108) |
| Auth | 48 of 1,354 (3.5%) | Clerk | 70.8% (34 of 48) |
| Payments | 84 of 1,354 (6.2%) | Stripe | 63.1% (53 of 84) |
Add the layers together and the picture gets narrower. Next.js plus Vercel was 326 of the 537 sites where we could see both a framework and a host (60.7%). Only 50 sites showed a framework, a database and a host at once. Among those 50, Next.js plus Supabase plus Vercel was 22 (44.0%), and Next.js plus Firebase plus Vercel was 8 (16.0%).
Which framework do indie startups use?
| Framework | Sites | Share of 778 detected |
|---|---|---|
| Next.js | 553 | 71.1% |
| Astro | 54 | 6.9% |
| WordPress | 31 | 4.0% |
| Framer | 27 | 3.5% |
| SvelteKit | 25 | 3.2% |
| Nuxt | 23 | 3.0% |
| Vue | 17 | 2.2% |
| Webflow | 14 | 1.8% |
| Angular | 11 | 1.4% |
| Ruby on Rails | 8 | 1.0% |
| Remix | 4 | 0.5% |
| Django | 3 | 0.4% |
| Other | 8 | 1.0% |
Next.js was the only framework above 10% in every group we looked at: 71.1% overall, 66.4% on launch sites (367 of 553) and 82.7% on revenue-making sites (186 of 225). The “other” row covers Wix and Shopify (2 each) and one site each on standalone React, Bubble, Squarespace and Vite. We found no Gatsby sites.
Where do indie startups host?
We named a host for 833 sites. Another 426 (31.5%) sat behind Cloudflare or CloudFront, and we couldn’t tell where the app ran. 95 sites (7.0%) gave no hosting signal.
| Host | Sites | Share of 833 named |
|---|---|---|
| Vercel | 438 | 52.6% |
| GitHub Pages | 75 | 9.0% |
| AWS | 72 | 8.6% |
| Framer | 40 | 4.8% |
| Netlify | 34 | 4.1% |
| Google Cloud | 25 | 3.0% |
| Webflow | 19 | 2.3% |
| Hetzner | 18 | 2.2% |
| DigitalOcean | 17 | 2.0% |
| Render | 14 | 1.7% |
| Fly.io | 13 | 1.6% |
| Firebase Hosting | 10 | 1.2% |
| Other named host | 58 | 7.0% |
AWS, Google Cloud, Hetzner and DigitalOcean come from a network-owner lookup (ASN) when a site’s DNS didn’t point at a known platform. That’s coarser than seeing “Vercel” or “Netlify” in the headers, so read those four rows as “the machine sits on this provider’s network.”
Supabase or Firebase? What indie startups use for databases
We could see a database on 108 sites, because most databases never talk to the browser. We counted only database or backend-as-a-service hostnames that show up in the browser.
| Database | Sites | Share of 108 detected |
|---|---|---|
| Supabase (Postgres) | 77 | 71.3% |
| Firestore / Firebase | 27 | 25.0% |
| Convex | 4 | 3.7% |
| Xata | 1 | 0.9% |
On the Supabase vs Firebase question, Supabase led by almost three to one among the sites that expose a database. Treat the Xata hit with caution, because it may be a false match.
A server-only Postgres database is invisible to this method. Neon, PlanetScale, Turso, Appwrite, PocketBase and MongoDB Atlas never showed up in our scan, even though founders list several of them on their TrustMRR profiles. Scanning page behavior instead of the homepage alone raised the count from 31 sites to 108, so the true picture is bigger than any of these tables.
Auth and payments: Clerk and Stripe
| Auth tool | Sites | Share of 48 detected |
|---|---|---|
| Clerk | 34 | 70.8% |
| Auth0 | 6 | 12.5% |
| Supabase Auth | 5 | 10.4% |
| Stytch | 1 | 2.1% |
| Kinde | 1 | 2.1% |
| Okta | 1 | 2.1% |
| Payment tool | Sites | Share of 84 detected |
|---|---|---|
| Stripe | 53 | 63.1% |
| Paddle | 15 | 17.9% |
| Lemon Squeezy | 10 | 11.9% |
| Polar | 6 | 7.1% |
Both are minimums. Stripe Checkout without Stripe’s browser script, and auth handled only on the server, don’t appear. Founders’ own profiles show how big the gap is for payments: 6,989 TrustMRR profiles name a payment tool, while our scan caught one on only 6.2% of sites.
Do sites that make revenue use a different stack?
We ran the same browser check on two groups. The launch sample is the live product sites from the Show HN and Product Hunt check behind our launch data report, and 971 of 976 loaded. The revenue sample is 383 of 386 websites from TrustMRR, a directory of startups with verified revenue.
| Signal | Launch sample | Revenue sample (TrustMRR) |
|---|---|---|
| Next.js among detected frameworks | 367 of 553 (66.4%) | 186 of 225 (82.7%) |
| Vercel among named hosts | 292 of 619 (47.2%) | 146 of 214 (68.2%) |
| GitHub Pages among named hosts | 74 of 619 (12.0%) | 1 of 214 (0.5%) |
| Supabase among detected databases | 36 of 59 (61.0%) | 41 of 49 (83.7%) |
| Clerk among detected auth | 19 of 32 (59.4%) | 15 of 16 (93.8%) |
| Stripe among detected payments | 38 of 55 (69.1%) | 15 of 29 (51.7%) |
| Next.js + Vercel among framework + host pairs | 215 of 384 (56.0%) | 111 of 153 (72.5%) |
Revenue-making sites concentrate more on Next.js, Vercel and Supabase, and GitHub Pages, a common home for early demos, almost disappears. The auth and payment rows rest on 16 and 29 revenue sites, so don’t lean on them. Nothing here shows the stack caused the revenue.
What founders say they use
TrustMRR founders fill in optional Frontend and Backend fields on their public profiles. We fetched 12,474 profile addresses from the site’s sitemap and parsed 10,256 of them (82.2%).
- Frontend: 5,455 profiles (53.2%) filled it in, and 5,212 of those (95.5%) named at least one known framework or native toolkit.
- Backend: 10,232 profiles (99.8%) filled it in. The field mixes databases, payments, languages and hosts, so it’s messy.
- Database: a recognizable database name appeared in 2,911 backends (28.4% of filled backends).
- Payments: a recognizable payment tool appeared in 6,989 profiles.
- Auth: a recognizable auth vendor appeared in only 170, so treat auth as thin here.
Mobile toolkits like Swift, React Native, Expo and Flutter show up in declared frontends and almost never in a website scan. That’s one reason founders’ declared stacks and our scan of TrustMRR websites don’t match.
What GitHub and developer surveys add
These measure different people, so we keep them separate and don’t blend them into one ranking.
GitHub, Show HN repos. We took 440 GitHub repos linked from Show HN posts (20 a month, January 2025 to October 2026). 97 (22.0%) had a root package.json. A mapped frontend framework showed up in 28 of them: React 13, Next.js 11, Vue 2, SvelteKit 1 and Astro 1. A data library appeared in 8: a generic Postgres driver 3, Prisma 2, Firebase 2, and one each for Turso, Drizzle and Supabase. No auth package mapped, and Stripe appeared once. Many Show HN repos are command-line tools or native apps, so this slice leans toward React libraries.
Stack Overflow Developer Survey 2025 (over 49,000 responses, 177 countries). These are developers worldwide, not indie startups.
| Question | Results |
|---|---|
| Web frameworks | Node.js 48.7%, React 44.7%, Next.js 20.8%, Astro 4.5%, Nuxt 4.0% |
| Databases | PostgreSQL 55.6%, MySQL 40.5%, SQLite 37.5%, Supabase 6.0%, Cloud Firestore 5.7% |
| Platforms and tools | AWS 43.3%, Cloudflare 20.1%, Vercel 10.6%, Netlify 5.9%, Supabase 5.4% |
State of JS 2025 (13,002 respondents, 24 September to 11 November 2025). Of people who answered each question, 83.6% had used React. Among meta-frameworks, 58.6% had used Next.js, 27.5% Astro, 23.6% Nuxt and 17.7% SvelteKit.
Wappalyzer, 2026. It measures the public web, not indie startups. Next.js was 43% of detected web frameworks (3.4 million sites), Vercel 21% of detected hosting platforms (2.3 million sites), and Stripe 21% of payment processors (1 million sites). Clerk was 3% of authentication tools and Auth0 2%.
What this means if you’re picking a stack
Popular doesn’t mean best. A founder picking Next.js, Vercel, Supabase, Clerk and Stripe is joining the most visible setup in this data. This data can’t say whether that setup helps products succeed faster.
If you go another way, you’d join a smaller group: Astro ran 54 sites, SvelteKit 25 and Nuxt 23, and 75 sites ran on GitHub Pages. We’re in the Astro group ourselves, hosting on Cloudflare. Once you’ve built something, our SaaS directories list shows where to launch it.
How we measured it
- The sites: 1,362 URLs: 976 from the launch sample plus 386 TrustMRR websites. 1,354 loaded in a real browser (99.4%), and 8 failed.
- The browser check: headless Chrome through Playwright on 10 November 2026, finished around 6:08 PM IST. It waited for the page, clicked and scrolled, and visited login, signup, app and dashboard URLs where they existed. It blocked images, media and fonts, and used a timeout of about 14 seconds.
- What it looked for: it read network requests, scanned up to six JavaScript chunks per site for database, auth and payment hostnames, checked runtime flags for Next, Nuxt, SvelteKit, Remix, Astro, Firebase and Supabase, looked up DNS records for each host, and ran an ASN lookup when the origin stayed unclear behind Cloudflare.
- Shares: we report a tool’s share only among sites where we detected that layer, and show coverage separately. An earlier pass counted undetected sites as a share, which diluted every real stack, so we redid it.
- Other sources: TrustMRR’s public profile pages, GitHub package.json files from Show HN, the Stack Overflow and State of JS 2025 results, and Wappalyzer’s public 2026 category tables.
We couldn’t read BuiltWith’s trend pages, and StackShare doesn’t publish a market-share table, so neither appears here.
Caveats
- Detected-only shares show which visible tool wins. They don’t say what share of all indie sites use it.
- Database, auth and payment coverage is low (8.0%, 3.5% and 6.2% of sites). A server-only database or server-only auth is invisible to a browser check.
- 31.5% of sites sat behind Cloudflare or CloudFront without a recovered origin. ASN labels name the network, not a product.
- Some matches can be false. Text matches (xata.io, prisma.io, firebaseapp.com, or a marketing mention of Lovable or v0) occasionally misfire, so treat the single Xata hit cautiously.
- TrustMRR fields are optional and inconsistent. Founders can list several tools, Backend often lists Stripe instead of a database, and 2,218 sitemap addresses didn’t parse as profiles.
- The launch sample leans toward survivors. It’s live sites from our earlier Show HN and Product Hunt check, and the Product Hunt part is a catalog add-on, not every launch.
- GitHub, Stack Overflow, State of JS and Wappalyzer cover small JavaScript repos, developers worldwide, survey takers and the whole public web, in that order. None is an indie-only sample.
FAQ
What tech stack do indie startups use?
Among the sites where we could see each layer, the most common stack is Next.js on Vercel with Supabase, Clerk and Stripe. Next.js ran 71.1% of detected frameworks, Vercel 52.6% of named hosts, Supabase 71.3% of detected databases, Clerk 70.8% of detected auth and Stripe 63.1% of detected payments.
Do most indie startups use Next.js?
Where we could detect a framework, yes: 553 of 778 sites (71.1%). We could detect a framework on only 57.5% of sites, so that isn’t 71% of all of them.
Is Supabase more popular than Firebase for indie startups?
Among the 108 sites that exposed a database, Supabase had 77 (71.3%) and Firebase or Firestore had 27 (25.0%). Server-only databases are invisible to our method, so the real split is unknown.
Where do indie startups host their sites?
Vercel hosted 438 of the 833 sites with a named host (52.6%). GitHub Pages came second at 9.0%, then AWS at 8.6%. Another 31.5% of all sites sat behind Cloudflare or CloudFront with no clear origin.
Do indie startups use Stripe?
Stripe held 53 of the 84 detected payment tools (63.1%), ahead of Paddle (17.9%), Lemon Squeezy (11.9%) and Polar (7.1%). We detected a payment tool on only 6.2% of sites, so that’s a minimum.
Cite this research
Indie Torch, “Indie Startup Tech Stack: What 1,354 Live Sites Run in 2026,” 10 November 2026. Browser checks of 1,354 live indie sites ran on 10 November 2026. Founder-declared stacks: TrustMRR public profiles.
Journalists and researchers can quote the figures with attribution. Please keep the caveat that these are shares among sites where a layer was visible, not shares of all indie sites.
Sources
- TrustMRR startup sitemap, TrustMRR
- TrustMRR llms.txt, TrustMRR
- Stack Overflow Developer Survey 2025: Technology, Stack Overflow
- State of JavaScript 2025, State of JS
- JavaScript frameworks, Wappalyzer
- Is Product Hunt worth it? 2025–2026 launch and failure data, Indie Torch
I'm Lourdes Paul Agilan, an indie maker and SEO consultant based in Chennai, India. I build small, focused software products and help B2B SaaS founders get found on Google and in AI answers. I like shipping fast, learning in public, and sharing what works along the way.