🔀 Splitpoint
A/B testing and feature flags you own forever.
Splitpoint at a glance
| Price | $49, once — no renewal |
| Replaces | VWO at $199/mo — roughly $2,388 a year, every year |
| Type | Self-hosted web app |
| Platforms | Your own server (VPS / Docker), any browser |
| Category | Utilities |
Splitpoint features
🧪 Experiments
Control plus N variants, weighted traffic split, holdout percentage, and URL/device targeting.
🎯 Goals two ways
A JS track() call, or automatic conversion when a visitor hits a matching pageview URL.
🪶 Tiny SDK
Vanilla JS under 3KB, no dependencies, never touches the DOM; assignments stick via deterministic hashing and localStorage so there is no flicker.
⚛️ React hook wrapper
An included hook for wiring code-level tests directly into your components.
📈 Results that mean something
Visitors, conversions and CR per variant with 95% CI, relative lift, a z-test p-value and a declare-winner workflow.
🚩 Feature-flag mode
Boolean flags with percentage rollout on the same deterministic hashing, for gradual releases with zero extra setup.
🚧 Mutual exclusion groups
Experiments sharing a group never run on the same visitor, so overlapping tests do not contaminate each other.
🧮 Honest counting
Goals only count for visitors actually assigned, and a DB unique constraint makes duplicate assignments impossible.
$49 once — that's the whole pricing page
VWO charges $199/mo — roughly $2,388 a year, every year. Splitpoint is a single payment, and it keeps working whether or not we ever sell another copy.
Or get Splitpoint inside the full suite — every app in the catalog, currently $97 for the bundle. See the bundle →
Splitpoint guides & pages
Honest answers
Is it really free on GitHub?
Yes — Splitpoint is MIT-licensed at github.com/bensblueprints/ab-testing-tool-mvp. The $49 buys the packaged Windows installer and updates; building from source is free.
Can I trust the statistics?
It runs a pooled two-proportion z-test and reports the two-tailed p-value, confidence and Wald 95% CI. Standard caveats apply: fix your sample size up front, do not peek-and-stop the moment a result looks significant, and remember 1 in 20 winners at 95% is noise.
Is this a subscription in disguise?
No. $49 once, no renewal, no per-visitor metering. Your experiment data lives on your own server.