Pre-launch · Validating demand

A/B test your landing pages without the PostHog tax.

Drop one small script. See which headline, hero, or CTA is converting - across every page in your portfolio. No JS framework. No session-replay bloat. No enterprise sales call.

Built by a solo founder running 6 landing pages who got tired of guessing.

dartlab.app/projects/my-portfolio

Mockup - this is what we're building. The dashboard is not live yet.

Why we exist

The current A/B testing market is split between enterprise tools you can't afford and OSS tools designed for product teams, not solo founders running multiple landing pages.

Tool What's frustrating
PostHogPowerful, but the cloud tier scales fast and self-hosting is a Kubernetes-grade lift.
Plausible / FathomBeautiful privacy-first analytics - but no A/B testing.
Vercel A/BOnly works on Vercel. No cross-page dashboard for your portfolio.
GrowthBookGreat OSS - but the UX assumes a product team with PMs and engineers, not a solo founder with 6 landing pages.
VWO / OptimizelyEnterprise pricing, sales-led onboarding. Not built for one-person operations.
KeakAI-autopilot hosted SaaS, $0.005 per impression. Built for marketers who want the AI to write the variants. Cost scales with traffic, lives outside your repo.
DartLabLives in your repo. Flat cost forever. You write the variants. One dashboard across your whole portfolio of landing pages.

How it works

Three steps. No SDK install, no schema design, no event-naming meetings.

1

Drop the script

One tag in your <head>. Works with any framework or plain HTML. ~6KB gzipped, async, doesn't block render.

2

Mark a winning event

Click on the button you care about (signup, purchase, sign-in) right in the dashboard. We'll track conversions to it. data-dart-goal if you prefer code.

3

Push variants, watch winners

Edit the headline / hero / CTA in the dashboard. Traffic splits 50/50. Bayesian credible intervals tell you when a variant has truly won - usually in days, not weeks.

Pricing

Two ways to be early.

Waitlist

Free

Get notified at launch

  • Email when we ship
  • Launch-day discount
  • No commitment
Join waitlist

Future public pricing (subject to change): $19/mo for 50K events · $49/mo for 250K events. Founders pay nothing more, ever.

The honest deal

Read this before you buy

  • The product is not built yet. This page is the validation experiment.
  • If 10+ founding seats sell by 2026-09-01, we start the build.
  • If they don't, we shelve the project and refund 100%. No questions.
  • Founding-seat buyers get the lifetime access promise only if we ship.
  • Waitlist is free with zero commitment - sign up if you want a "tell me if this happens" ping.

Why open about it? Because the indie hacker community has been burned enough by vaporware. The honest pre-sell is the validation. If you don't trust it'll get built, the waitlist exists for exactly that reason.

Why I'm building this

I run a portfolio of 6 small SaaS / digital-product "darts" - each one a self-contained landing page or app I'm trying to find product-market fit for.

Last quarter I had no idea which headline on bidcalc.dev was actually converting. I tried PostHog (too much setup), looked at Vercel A/B (only works inside Vercel), and almost installed GrowthBook before realizing it would take a week to learn for a portfolio of 6 pages I check casually.

So I started spec'ing what I actually wanted: one script, one dashboard across all my pages, just A/B. If 10 other people running landing pages pre-buy this for $99 lifetime, I'll build it. If they don't, my portfolio doesn't need it badly enough to justify the build either.

Either way, you'll know in two weeks.

Get notified at launch

Free. Zero commitment. Ping when there's news.

FAQ

No. This is pre-launch validation. If 10+ founding seats sell by September 1, 2026, the build starts. If not, the project is shelved and all founding-seat buyers are refunded.
Full refund, no questions asked. The point of the pre-sell is to validate demand - if it doesn't validate, no one should be out money.
Realistically 6-10 weeks from validation. The first version will cover script ingestion, two-variant A/B tests, conversion goals, and a single dashboard for your whole portfolio. No session replay, no funnels, no heatmaps in v1.
Bayesian credible intervals (95% threshold by default). Frequentist p-values get the headline in textbooks but Bayesian methods make decision-making cleaner for the small-traffic landing pages this is built for.
Target is ~6KB gzipped, async, no render-blocking. Cached on a global CDN. If it's slower than that at launch, founding-seat buyers can refund.
Yes. Plain HTML, any framework, any host. The script is framework-agnostic.
No PII collected by default - just anonymous variant assignment + conversion event. Cookieless mode planned for v1. EU users will get a data-residency option once we have the volume to justify EU hosting.
Yes - that's the whole point of "founding." If we ship and add more usage, founders never pay more, even if we add new pricing tiers later.