Find a better headline

Continuously test your landing page copy to find what converts best.

yoursite.com

Your landing page writes itself

Start building
Austin, USA🇺🇸
Headlines tested▲ 21% vs. baseline
Your landing page writes itselflive4.7%
Ship your idea this weekend4.5%
Stop guessing which headline works4.5%
Turn visitors into customersbaseline3.9%
Buttons tested▲ 15% vs. baseline
Get started free4.7%
Start buildinglive4.7%
Try it nowbaseline4.1%
Week 1Week 2Week 3Week 4Week 5Week 6

Test every piece of copy and always show the most optimal one.

See what's winning

Leading version

Your landing page writes itself

4.7% conversion rate

Lift

▲ +20%

Leading version vs original

Visitors

8,700

Across all versions

Conversions

359

Version
  1. Your landing page writes itselfWinning2,9601394.7%+20%−10%+50%
  2. Ship your idea this weekend1,210554.5%+16%−21%+54%
  3. Stop guessing which headline works980444.5%+15%−25%+55%
  4. Turn visitors into customersControl1,840723.9%—
  5. Copy that converts, automatically760273.6%−9%−50%+31%
  6. The smarter way to write headlines540152.8%−29%−71%+13%
  7. AI-powered A/B testing for everyone41071.7%−56%−96%−17%
  • The copy converting best right now
  • How much better it is than your original
  • Your goal: signups, checkouts, demos booked…
  • Bar clear of the line? A real win, not luck.
  • Winners get shown to more visitors
  • Your original copy. Everything is measured against it.
  • Losers get less and less traffic
  • Add a new version any time. It joins the test right away.

How it works

It can take weeks or even months to test all the copy on our landing page. And while that is happening, you are losing potential customers to your competitors.

Splitduck works differently.

  • Instead of testing one headline at a time, it continuously tests all the copy on your landing page and automatically shows the best performing version to your visitors.
  • Initially, it picks a version at random, but as soon as it gets some data to prove that one version is better than the others, it will show that version to more visitors.
  • Over time, it will converge on the best performing copy.

Set it up your way

Do it yourself

  1. 1

    Add the script to your site

    <script src="https://splitduck.com/splitduck.js" …>
  2. 2

    Tag the text you want to test

    <h1 data-splitduck="headline">…</h1>
  3. 3

    Track your goal

    splitduck.track("signup_completed")
Read the setup docs →

Let your agent do it

Using Claude Code, Cursor, or another coding agent? Paste this into it and it will install the script, tag your copy, and track your goals.

Set up SplitDuck on this site by following the instructions at https://splitduck.com/agent.txt

Read the instructions yourself at splitduck.com/agent.txt

FAQ

What's the catch with free?

No catch. We're early and we'd rather have you using it and telling us what to build. You get unlimited sites and tests with no card required. Fair use applies to traffic, and we plan to add paid plans in the future.

Start free
I'm not a developer. Can I still use this?

Yes, if you use an AI coding agent like Claude Code or Cursor. Paste this in and it adds the script, tags your copy and tracks your signups for you:

Set up SplitDuck on this site by following the instructions at https://splitduck.com/agent.txt

Without an agent, someone who can edit your site's HTML can do it in a few minutes.

How long until I see results?

It starts learning from your very first visitors. From the start, most of your traffic sees whichever version is converting best so far, and that shifts as the data comes in. No waiting weeks for a test to finish before you benefit.

I've tried A/B testing. It took forever and nothing won.

That's the problem SplitDuck fixes. Classic A/B tests run one change at a time and split traffic evenly until there's a verdict, so you lose visitors to the weaker copy the whole time. SplitDuck tests every version at once and keeps sending more visitors to the winner while it learns. See the results table above for how that looks.

Is it just about finding the best headline?

It's the fastest way to learn what your customers actually care about. Every version you test is a different pitch, and the winner tells you which one resonates. And it works on any text: buttons, subheadlines, body copy.

Will it slow down or break my site?

No. The script is small and loads asynchronously, so it doesn't block your page. Tagged text stays hidden for a moment while a version is picked, and if anything goes wrong, your original copy shows within half a second.

Who is SplitDuck for?

Early-stage teams with a live product and a landing page: founders, product owners and whoever runs growth. If you're looking for product-market fit, or want more visitors to reach your signup or checkout, it's built for you.

What counts as a conversion?

Whatever action you want more of: a signup, a started checkout, a booked demo. You track it with one line of JavaScript and pick it as the goal for each experiment.

Still reading? Just try it.

It's free, and setup takes minutes.

Get started free