Skip to content

You ship.
Dispatch tests.

Dispatch writes, fixes and maintains Patrol E2E tests in your repo. Autonomously.

Watch the demoGet the first month free
  1. Ship the PR
  2. Dispatch tests it
  3. Review & merge

Your last month, roughly

40features implemented
0E2E tests written

Most of it written with AI. Nobody had the hours to prove it works. So the bug ships, and a developer loses half a day to it.

One feature in.
One E2E test out.

  1. Ship the PR
  2. Dispatch tests it
  3. Review & merge
Open feat: add promo codes to checkout #482
anna wants to merge feature/promo-codes into main · 14 files changed
  1. anna added the patrol:cover label
  2. patrol-dispatchbot

    On it. Booting a device and reading #482.

  3. Device ready · Pixel 8 · Android 150:48
    Wrote 3 tests in integration_test/checkout/4:10
    1 failed on its own setup · fixed, rerun8:31
    3 passed · recording saved11:42
  4. patrol-dispatchbot

    Draft #483 is ready: 3 tests, 2 files, +214, recording attached.

    Draft test: cover promo codes at checkout #483
    integration_test/checkout/promo_code_test.dart · promo_code_limits_test.dart
    ▶ Run recording · 0:58
    For a humanA 100% promo code leaves the total at $0.01. Written up in the draft. App code untouched.

Test files only. A bug it finds is written up for a human, not patched behind your back. Dispatch never merges.

Watch the demo

A silent walkthrough: you open a PR, Patrol Dispatch works out what to test from your diff, writes and runs the test on an emulator, and you get a draft PR with the verified test.

10× the QA.
At a fraction of the cost.

Compared onHiring QAPatrol Dispatch
First testin a few weekswithin an hour
Coveragea few flows a weekevery PR you label
Hoursnine to five3 a.m. included
Costa salary and a device farma fraction of that
Setuphiringnothing to host
You keepone person's memorytests in your repo

You don't have to read the tests. Every draft comes with a recording of them passing. Sixty seconds, and you know the feature works.

Why Dispatch.

You own your tests

Real Patrol tests, committed to your repo. They stay yours if you cancel.

Deterministic result, every run

Once written, the tests run in your CI with no AI involved. The same test gives the same answer today and in a year.

Nothing to host

We bring the session and the devices. Your team ships a feature and reviews a draft.

Built by the Patrol team

The E2E framework Flutter already tests with, written by the people who maintain it.

Built on the framework
Flutter already trusts.

>700k
Patrol downloads in the last 30 days
1.4k
GitHub stars on leancodepl/patrol
726
pub.dev likes
2022
First Patrol release. Four years of native Flutter testing since.

Before you ask.

We don't have Patrol yet.
Fine. A Flutter app on GitHub and a CI build is enough. Adding Patrol is part of the first call.
We have a small QA team.
Point Dispatch at the backlog. QA keeps the exploratory work only a human can do.
Real tests, or an AI clicking through the app?
Real Patrol tests in your repo. No model at run time, the same result every run.
How much?
Per repository, an order of magnitude below a QA hire. Founding teams get the first month free. Plans on the call.

Founding teams

First month on us.

A human replies with a calendar link. On the first call we connect one repository, and the first draft PR lands before the call ends.

  • A LeanCode engineer sets it up on your repo
  • Every draft reviewed before you see it
  • No QA hire. No test-writing sprint.