> ## Documentation Index
> Fetch the complete documentation index at: https://docs.convonexai.tech/llms.txt
> Use this file to discover all available pages before exploring further.

# Your app's own tests

> The tests the builder writes with your app's code: what they prove, how strong they are, and how to run them yourself.

Your app comes with its own test suite. The builder writes the tests while
it writes the code, and each test names the rule it proves. Creator runs the
suite every time a story is verified, and shows you what it said.

The suite is part of your app. It is in your code, it goes with the code
when you export it, and your own developers can run it.

## What a test proves

Every story has rules: what the app must refuse, what it must limit, who may
do what, what happens to a record after an action. Each rule has a short id.
A test carries the id of its rule in its name:

```ts theme={null}
it("[c-b0d13c292e] a deadline of yesterday is refused", () => { … });
```

That is how Creator knows which rule a test is about. A test with no id
still runs and is counted. It proves no rule.

## The four states of a rule

| State | Meaning |
| - | - |
| **Proved by the app's tests** | A test that names this rule ran, and every one that ran passed. |
| **Did not work** | A test that names this rule failed. |
| **Not tested** | No test names this rule, or the builder skipped the test and said why. |
| **Could not run** | Tests name this rule, and Creator could not run them. This says nothing about your app. |

When the builder cannot test a rule, it says so in the suite, with the
reason. You see the reason beside the rule.

## Test strength

A test that can never fail proves nothing. To find those, Creator makes
small deliberate changes to the code a story wrote, one at a time, and runs
the story's tests again. It reverses a comparison, moves a limit, swaps an
"and" for an "or". The changes are made in a copy; your app and your preview
are never touched.

| You see | Meaning |
| - | - |
| **Tests: strong** | The tests noticed at least four in five of the changes. |
| **Tests: mixed** | They noticed at least half. |
| **Tests may not catch a break** | They noticed fewer than half. |
| **Test strength not measured** | There was nothing to measure, and the reason is given. |

Test strength is measured a few minutes after a story is verified, so it
appears a little later than the counts. It never changes whether a story is
verified or built.

## Where you see it

| Place | What it shows |
| - | - |
| Story card | How many of the story's tests pass, and the test strength. |
| Story page | Each rule, the tests that name it, and their results. |
| [Tests tab](/verification/tests-tab) | The whole suite, by story and by rule. Tests that name no rule are listed last. |
| Publish screen | How many rules the tests proved, how many did not work, how many were not tested. |

The line on the publish screen is information. It does not decide whether
you can publish.

## Tests that read or write data

Some tests need a database. They get one of their own, emptied before every
run and filled with the app's own starting records. Your app's data is never
used by a test.

If that database cannot be made ready, the tests that need it read "could
not run", and Creator says why.

## Running the tests yourself

The Tests tab shows the command. For most apps it is:

```bash theme={null}
npm test
```

## Quick builds

A Quick build is not asked to write tests. If it has tests anyway, Creator
runs and counts them.