# Kotlin coverage: Kover, JaCoCo, and Supercov

> Compare Kover, JaCoCo and Supercov for Kotlin: Gradle reports and verification, per-test coverage for Kotest and Spock, MC/DC and assertion coverage.

Web page: https://supercov.com/compare/kotlin-code-coverage · Published 2026-09-25
All Supercov pages for agents: https://supercov.com/llms.txt · Full text: https://supercov.com/llms-full.txt

## Kover vs JaCoCo vs Supercov

Start with Kover if your Kotlin project builds with Gradle: it is the Kotlin
team's own coverage plugin. JaCoCo is the long-standing JVM option. Supercov measures per test and adds MC/DC and
assertion coverage.

| Feature | Kover 0.9.9 | JaCoCo 0.8.15 | Supercov 2.0.0 |
|---|---|---|---|
| Code quality scoring | × Not built in | × Not built in | ✓ Supported |
| Security review | × Not built in | × Not built in | ✓ Supported |
| Line coverage | ✓ Supported | ✓ Supported | ✓ Supported |
| Branch coverage | ✓ Supported | ✓ Supported | ✓ Supported |
| MC/DC | × Not built in | × Not built in | ✓ Supported |
| Per-test coverage | × Not built in | × Not built in | ✓ Supported |
| Assertion coverage | × Not built in | × Not built in | ✓ Supported |
| Reports | ✓ Supported | ✓ Supported | ✓ Supported |

Standard coverage workflows in the versions shown; custom plugins are outside this table. Supercov measures Kotlin through the same JVM path as Java, taking test identity from the framework’s own lifecycle, so Kotest and Spock tests keep the names they report. A condition that also narrows a type carries no MC/DC obligation, and the run names every such case. Kover measures with its own coverage library by default and can switch to JaCoCo. Both report line and branch coverage per run; neither computes MC/DC or per-test coverage. Per-test coverage means individual tests, not just test files. Supercov needs a supported runner. Assertion coverage links source to checked properties through agent review; it is not a mutation score or a correctness guarantee.

## Kover: Gradle tasks and its own agent

Kover instruments bytecode with its own coverage library by default. The
[Gradle plugin](https://kotlin.github.io/kotlinx-kover/gradle-plugin/) reports
lines by default, and instructions or branches when you ask:

```sh
./gradlew koverHtmlReport
./gradlew koverXmlReport
./gradlew koverVerify
```

The XML report is JaCoCo-compatible, so CI tools that read JaCoCo read it too,
and `koverVerify` fails the build below the bounds you configure. `useJacoco()`
switches the engine to JaCoCo, with fewer filters available. Either way the
report covers the run, not each test.

## JaCoCo on Kotlin

JaCoCo's Gradle plugin works on Kotlin as on Java:

```sh
./gradlew test jacocoTestReport
```

It counts bytecode, and its
[counters](https://www.jacoco.org/jacoco/trunk/doc/counters.html) are the same
as for Java: instructions, branches, lines, methods and classes.
Neither JaCoCo nor Kover computes MC/DC or per-test coverage.

## What Supercov adds

With JDK 17 or newer and Gradle:

```sh
brew install supercorp-ai/tap/supercov
supercov -- ./gradlew test
supercov runs latest gaps
```

Supercov measures Kotlin through the same JVM path as Java. It takes test
identity from the framework's own lifecycle, so Kotest and Spock tests keep the
names they report, and each covered line is tied to the tests that ran it.

For [MC/DC](https://supercov.com/docs/mcdc.md) it observes each condition in a decision. The one
exception is a condition the compiler also reads: `x is String` or `x != null`
narrows the type for the code beneath it, and wrapping it to watch its operands
would stop the code compiling. Supercov leaves such an `if` as written, records
which way it went from inside its arms, and gives it no MC/DC obligation. The
run names every such case instead of skipping it silently.
[Assertion coverage](https://supercov.com/docs/assertions.md) then shows which statements a passing
test actually checks. The [Kotlin guide](https://supercov.com/docs/kotlin.md) has the setup and a
runnable example.

Keep Kover for its Gradle integration and reports. Use Supercov when you, or
your coding agent, need per-test evidence and the missing conditions and
checks named. None of these scores proves the program correct.


## Feature sources

- [Kover: Gradle plugin and reports](https://github.com/Kotlin/kotlinx-kover/blob/v0.9.9/README.md)
- [Kover: Measured metrics](https://kotlin.github.io/kotlinx-kover/gradle-plugin/)
- [JaCoCo: Coverage counters](https://www.jacoco.org/jacoco/trunk/doc/counters.html)
- [JaCoCo: Agent and reports](https://www.jacoco.org/jacoco/trunk/doc/agent.html)
- [JaCoCo: HTML report](https://www.jacoco.org/jacoco/trunk/doc/cli.html)
- [Supercov: Coverage model](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/coverage-model.md)
- [Supercov: Supported runners](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/supported-suites.md)
- [Supercov: Assertion coverage](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/assertions.md)
- [Supercov: Code quality](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/quality.md)
- [Supercov: Security](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/security.md)
- [Supercov: Reports](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/reports.md)
- [Supercov: CLI reference](https://github.com/supercorp-ai/supercov/blob/v2.0.0/docs/cli.md)