Skip to main content
Fleack is an A/B testing platform for mobile apps that works entirely at the network layer. Instead of shipping a new SDK version or waiting on a store review, you point one DNS record at Fleack and your team can start testing prices, reward amounts, ad frequencies, feature toggles, and copy — across every active app version, instantly. This page explains what Fleack is, who it’s built for, and where it fits in your testing stack.

​
How Fleack works

Fleack sits between your mobile app and your backend API. When your app calls api.your-game.com, the request flows through Fleack first. Fleack forwards it to your real backend unchanged, receives the response, and — when a test is running — rewrites the relevant parameter before returning it to the app. Your app sees a normal HTTP response. Your backend doesn’t change. The only infrastructure touch is a single CNAME record. Three capabilities make this practical for growth teams:
  • AI lever discovery. Fleack captures your endpoint responses and uses a GPT-class model to identify which parameters are worth testing — reward values, prices, frequencies, toggle flags — without any manual JSON spelunking.
  • No-SDK integration. There’s nothing to install in your mobile app. Tests reach users on every app version already in the field, including versions you can no longer update.
  • Bayesian results. Instead of waiting for p-value thresholds, Fleack computes a win probability per variant in real time. When a variant hits ≥ 90% probability of beating the control, you can promote it to 100% of traffic with one click.

​
Who Fleack is for

Fleack is built for mobile growth teams — product managers, marketers, and growth engineers who want to move fast without filing a ticket every time they want to test a number. If your team currently coordinates with engineering just to change an ad frequency or a daily reward amount, Fleack removes that dependency. You discover the lever in the backoffice, set the variants, and launch. Engineering never gets paged.

​
When Fleack is not the right tool

Fleack requires that your app’s backend is reachable at a public hostname you control and can set a CNAME on. Two situations where it won’t work:
  • Bundled configuration. If your app reads config from a file compiled into the binary (rather than a live API call), Fleack has no response to intercept.
  • Offline or flag-evaluation SDKs. If you need flags evaluated on-device without a network round-trip — for offline mode or sub-millisecond latency requirements — Fleack is not designed for that use case.
If either of those applies, Fleack isn’t the right fit, and that’s fine.

​
Where to go next

Quickstart

Set up DNS, send traffic, and launch your first test in five minutes.

Core concepts

Understand endpoints, levers, tests, segments, and the Bayesian results engine.

How integration works

DNS setup, fail-open behaviour, latency budget, and traffic flow details.

Recipe: interstitial frequency

A concrete end-to-end walkthrough of a common mobile growth test.