# How Testifly Works

Testifly tests your web app in four stages. First, AI agents **discover** your app and describe its features. Then they **write tests** in plain language. Next, they **automate** each test, which means they turn it into test code. Finally, Testifly **runs** the code and reports the results. This page explains the words Testifly uses and what each status means.

## What are the building blocks of a project?

| Term | What it is |
| --- | --- |
| **Project** | One web app you want to test. It has its own settings, users, features and tests. |
| **Environment** | One copy of your app, for example staging or QA. Each environment has its own base URL, test users and settings. Every project starts with **Main Environment**. |
| **Feature** | One testable part of your app, usually a page or an area of a page. A feature has a URL, an **Overview**, a detailed **Definition** and a list of UI components. Features can contain sub-features. |
| **UI component** | One element on a page, such as a button, a field, a menu or a table. Testifly stores a name, a description and a screenshot of each component. Test steps refer to components by name. |
| **Test** | A check of one behaviour. It has a title, a description and numbered steps. Each step has an action and an expected result. |
| **Prerequisite** | A test that must pass before another test can start, for example “Page Loaded Verification”. Testifly always runs prerequisites first. |
| **Test code** | The automated version of a test. Testifly writes it for you. You can read it on the test page. |
| **Test suite** | A named group of tests that you run together, for example **Sanity**, **Smoke** or **Regression**. Suites can run on a schedule. |
| **Run** | One execution of a set of tests. Each run records the result, a video and a timeline for every test. |
| **Job** | A piece of background work, such as feature discovery, automation or a run. You can follow jobs on the **Jobs** page. |
| **Runner** | The machine that opens a browser and runs the tests. Testifly provides cloud runners. For apps behind a VPN, you install a [Remote Runner](/content/documentation/remote-runners/index.html) in your own network. |

## How does work flow through Testifly?

1. **Discover.** An agent visits your app, finds its pages and creates a feature for each one. It records the UI components and clicks on them to learn what they do.
2. **Write tests.** For each feature, the agent writes tests in plain language. Tests fall into four categories: **Visibility**, **Functionality**, **Business Logic** and **Negative** (checks that the app handles wrong input correctly). The agent also places tests into your test suites, based on each suite’s name and description.
3. **Automate.** For each test, an agent opens the page in a real browser, follows the steps one by one and writes the code for each step. It checks that the code works before it moves on.
4. **Run.** Testifly executes the code, records a video and saves the result. If a step fails because the page changed, Testifly tries to repair the test by itself.
5. **Ask you.** When Testifly cannot tell whether a failure is a change in your app or a bug, it marks the test **Awaiting Your Input** and asks you.

You can also skip discovery and write features and tests yourself. The later stages are the same.

## What do the test statuses mean?

Every test has one status. The dashboard groups the statuses into two families.

### Automation statuses

These statuses describe the test before it has working code, or why it could not get code.

| Status | What it means | What you can do |
| --- | --- | --- |
| **Draft** | The test is saved but not sent for automation. A test also returns to Draft when you change its description or steps. | Review the steps and click **Automate test**. |
| **Automating** | An agent is writing the code for this test now, or the test is waiting in line. | Wait. You get an email when automation finishes. |
| **Automation Failed** | The agent could not write working code, for example because a step was unclear or the process took too long. | Read the error on the test page, improve the steps and click **Automate test**. |
| **Product Bug Detected** | The agent found behaviour that looks like a real bug in your app. | Read the explanation. Fix the bug, or re-automate the test if you think the agent was wrong. |
| **Not Applicable** | The test cannot work as written, for reasons outside the test itself. A common cause is missing data, for example an empty list. | Add the missing data or delete the test. |
| **Authentication Failed** | The agent could not log in to your app. | Check the test user in **Settings → Environments**. |
| **Faulty Prerequisite** | A prerequisite of this test does not work, so this test cannot start. | Fix or re-automate the prerequisite test first. |
| **Excluded Path** | The test reached a URL that you excluded in **Settings → Exclusions**. | Change the test or the exclusion list. |
| **Over plan limit** | Your project reached the test limit of its current plan, so Testifly could not automate this test. | Choose a bigger plan in **Settings → Billing**, or delete tests you no longer need. Then click **Automate test**. |

### Run statuses

These statuses describe a test that has working code.

| Status | What it means |
| --- | --- |
| **Ready To Run** | The test has code and is waiting for its next run. |
| **Run Scheduled** | The test is part of a run that is about to start. |
| **Running** | The test is running now. |
| **Test Run Passed** | The last run passed. |
| **Test Run Failed** | The last run failed. Open the run to see the video and the error. |
| **Test Run Skipped** | The test did not start, for example because the login or a prerequisite failed, or a bot check blocked the browser. |
| **Awaiting Your Input** | The last run failed and Testifly needs your answer to continue. See [Runs and Results](/content/documentation/runs-and-results#what-does-awaiting-your-input-mean/index.html). |

### Run results

A whole run gets one of three results: **Pass** (all tests passed), **Fail** (all tests failed) or **Partial Pass** (some passed and some did not).

## Which tests does Testifly run?

A test must have code to run. Tests in **Draft**, **Automating**, **Automation Failed** or **Over plan limit** have no working code, so they never run.

When you run a **whole feature** or a **test suite**, Testifly also leaves out:

- Tests marked **Product Bug Detected** or **Not Applicable**. Their status stays in place so the problem stays visible until you act on it.
- Tests that are **Awaiting Your Input**. Answer the question on the run page first.

When you pick **specific tests** – with **Run test** on a test page or **Run filtered tests** on the features list – every chosen test that has code runs, whatever its status.

## Where do I go next?

- [Features](/content/documentation/features/index.html) – add, discover and manage features.
- [Tests](/content/documentation/tests/index.html) – write, edit and automate tests.
- [Runs and Results](/content/documentation/runs-and-results/index.html) – run tests and read the results.