TestProject Free Test Automation: Complete Guide

TestProject Free Test Automation: Complete Guide

TestProject was a cloud-based, free-forever test automation platform built on top of Selenium and Appium. It wrapped the complexity of those frameworks behind a collaborative dashboard, an addon ecosystem, and an AI-powered recorder — all at no cost. Tricentis acquired TestProject in 2021 and sunset it in 2023, but understanding what it offered remains valuable: it defined a template for what modern free test automation should look like, and many teams are now searching for equivalent alternatives.

This guide covers the complete TestProject feature set, how it worked under the hood, and what to look for in a replacement.

What TestProject Was

TestProject provided three things simultaneously:

  1. A hosted agent — a lightweight binary you installed on your machine or CI server that executed tests locally but reported results to a cloud dashboard.
  2. An OpenSDK — open-source Python, Java, C#, and Ruby SDKs that extended Selenium WebDriver and Appium with built-in reporting, self-healing element detection, and addon support.
  3. A no-code recorder — a Chrome extension that captured browser interactions and generated test steps without writing a line of code.

The free tier had no time limit and no test count cap. You needed only a TestProject account (free) and a machine to run the agent on.

The Free Forever Plan

The free plan included:

  • Unlimited tests and test steps
  • Unlimited local agent executions
  • Cloud dashboard for reports and analytics
  • Addon marketplace access (1,000+ community addons)
  • Team collaboration (up to 3 users)
  • RESTful API access for CI/CD integration
  • Mobile testing via Appium (Android and iOS)

The paid plans added parallel execution on TestProject's cloud infrastructure and increased team seats, but the core automation capability was entirely free.

How the OpenSDK Wrapped Selenium

When you installed the TestProject SDK, your tests ran through a proxy driver that sat between your test code and the real browser driver:

# Standard Selenium
from selenium import webdriver
driver = webdriver.Chrome()

# TestProject SDK — same API, added reporting
from src.testproject.sdk.drivers import webdriver
driver = webdriver.Chrome()
# All Selenium methods work identically
driver.get("https://example.com")
driver.find_element(By.ID, "username").send_keys("testuser")

The proxy driver automatically:

  • Reported each step (pass/fail, screenshot) to the cloud dashboard
  • Applied smart element locator strategies to find elements even when attributes changed
  • Collected timing data for performance analysis

For Appium mobile testing, the pattern was identical:

from src.testproject.sdk.drivers import webdriver

desired_caps = {
    "platformName": "Android",
    "deviceName": "emulator-5554",
    "app": "/path/to/app.apk"
}
driver = webdriver.Remote(
    "http://localhost:4723/wd/hub",
    desired_caps
)

The same cloud reporting applied to mobile sessions automatically.

The Addon Ecosystem

TestProject's addon marketplace was its most distinctive feature. Addons were packaged test actions contributed by the community — reusable blocks you could drag into no-code tests or call from SDK tests.

Common addon categories:

  • Web interactions: File upload helpers, iframe handlers, shadow DOM navigators
  • Validations: JSON response validators, regex matchers, database query checkers
  • Integrations: Jira ticket creators, Slack notifiers, email senders
  • Utilities: Random data generators, date calculators, CSV readers

In the OpenSDK, addons were called as typed action objects:

// Java SDK — using a community addon
ClearFieldsAction clearAction = new ClearFieldsAction();
driver.addons().execute(clearAction, By.cssSelector("form"));

AI-Powered Test Recording

The TestProject recorder used AI to infer element semantics beyond raw XPath. When you recorded a click on a button labeled "Submit Order," the recorder stored multiple locator strategies ranked by stability:

  1. Text content: //button[text()='Submit Order']
  2. Aria label: //button[@aria-label='submit-order']
  3. CSS class + position fallback

During playback, if the primary locator failed, the agent tried the next strategy automatically — this was the self-healing behavior that TestProject marketed as a key differentiator.

Test Scheduling and Reporting

From the dashboard you could schedule any test or job (test suite) to run on a cron schedule. The agent on your machine polled for scheduled jobs and executed them locally, so you never paid for cloud compute.

Reports included:

  • Step-by-step pass/fail with screenshots at each step
  • Execution timeline
  • Flaky test detection (steps that pass/fail inconsistently)
  • Trend graphs over time

Limitations You Had to Work Around

Agent dependency: Every execution required the TestProject agent running on a machine you controlled. There was no serverless execution on the free plan — you provisioned the infrastructure.

Agent setup per environment: Each CI worker needed the agent installed and authenticated. In ephemeral Docker-based CI, this meant building a custom image with the agent pre-installed.

No built-in cross-browser parallel execution: The free plan ran tests sequentially on whichever browsers the local agent had installed. Parallel execution required the paid cloud plan.

Addon version pinning: Tests tied to community addons could break when addon authors released updates. There was no package lock equivalent for addons.

What TestProject Meant for the Market

TestProject proved that free, production-grade test automation was viable — and that teams would adopt it at scale when the barrier to entry was near zero. Over 200,000 teams used the platform before shutdown. The model it pioneered (free core automation + paid cloud execution) is now the baseline expectation.

The No-Infrastructure Alternative

TestProject required you to run an agent, manage the machine it ran on, and deal with driver version compatibility (ChromeDriver versions must match Chrome versions exactly). For teams that wanted automation without those operational responsibilities, platforms like HelpMeTest take a different approach.

HelpMeTest is a cloud-hosted SaaS where tests run in plain English — no SDK, no agent, no driver management. A 14-day free trial gives you 24/7 monitoring at 5-minute intervals and full CI/CD integration. There's no infrastructure to provision. You write what you want to test; the platform handles execution.

For teams evaluating TestProject replacements, the key question is whether you want to own the execution layer (agent-based model) or hand it off entirely (cloud-hosted model). Both are valid — the choice depends on how much control versus convenience your team needs.


Key takeaways:

  • TestProject wrapped Selenium/Appium with cloud reporting and self-healing locators, all free
  • The addon ecosystem extended test capabilities without writing custom code
  • The agent model gave full control but required infrastructure management
  • Cloud-hosted alternatives like HelpMeTest eliminate agent setup at the cost of some control

Read more

Start now free