What is it?

Xcode-like HTML report for Unit and UI Tests

▶ Open a live report — rendered from this repository's own sample test run, republished on every merge to main. Past releases are at /v/.

Features

• Supports parallel testing
• Supports attachments:
  • .png
  • .jpeg
  • .heic
  • .txt
  • .log
  • .mp4
  • .gif
• Navigate through the report with the keyboard's arrow keys
• Filter out successful, failed, skipped, or mixed-result tests
• Displays information about the target device
• Displays activity logs
• Junit report(-j flag)
• Json report(--json flag)
• Shrink bundle size by removing unattached files
• Automatically convert heic images to browser-friendly format
• Render as a single html file with inline attachments or as a bundle
• Downsize image attachments

Installation

Homebrew (recommended)

Install via Homebrew

Install latest stable version

brew install xctesthtmlreport

Install latest from main branch

brew install xctesthtmlreport --HEAD

Mint

Install via Mint

Install latest stable version

mint install XCTestHTMLReport/XCTestHTMLReport

Install latest from main branch

mint install XCTestHTMLReport/XCTestHTMLReport@main

Usage

Run your UI tests using xcodebuild without forgetting to specify the resultBundlePath

$ xcodebuild test -workspace XCTestHTMLReport.xcworkspace -scheme SampleApp -destination 'platform=iOS Simulator,name=iPhone 14,OS=16.0' -resultBundlePath TestResults

Then use the previously downloaded xchtmlreport tool to create the HTML report. Additionally, -i flag is also available to inline all resources, this is convenient for exporting the html file standalone. HTML file will be much heavier but much more portable.

$ xchtmlreport TestResults.xcresult

Report successfully created at ./index.html

Multiple Result Bundle Path

$ xchtmlreport TestResults1 TestResults2

Report successfully created at ./index.html

This will create only one HTML Report in the path you passed with the -r option

Progress

A large result bundle takes a while, and a silent run is hard to tell from a hung one. When standard error is a terminal, each phase is reported as it finishes, with the time it took:

$ xchtmlreport TestResults.xcresult
    Exporting 2531 attachments          4.8s
  Reading TestResults.xcresult         18.2s
  Rendering 1284 tests                 12.4s
  Writing report                        0.9s
Wrote ./index.html (412 MB)            36.3s

Report successfully created at ./index.html

Those timings also answer "which part is slow?" on your own bundle rather than in the abstract. Indented lines are nested inside the line below them — attachment export happens during the read.

The rules are git's. Progress goes to standard error, so a pipeline reading the report path off stdout (xchtmlreport … | xargs open) is unaffected. It is off when standard error is not a terminal, so scripts and CI logs are unchanged. --progress forces it on anyway, and --quiet turns it off.

Generate Junit Reports

You can generate junit reports with the -j flag

$ xchtmlreport -j TestResults1

Report successfully created at ./index.html

JUnit report successfully created at report.junit

Generate JSON Reports

You can generate json reports with the --json flag

$ xchtmlreport --json TestResults1

Report successfully created at ./index.html

JSON report successfully created at ./report.json

Starting in 4.0, report.json is our own documented, versioned schema — docs/json-schema.md is the contract. The schema is identical whichever result reader produced it: same keys, same nesting, same schemaVersion. A few values legitimately differ between readers — the four differences listed under "Choosing the result reader" below, plus testCase.arguments, which only the modern reader can populate — and the contract documents each one. Earlier versions dumped xcresulttool's legacy object graph verbatim; that graph is Apple's internal shape and disappears together with the legacy commands, so 4.0 replaces it once, deliberately. The change is visible at a glance — before:

[{"_type":{"_name":"ActionsInvocationRecord"},"actions":{"_type":{"_name":"Array"},"_values":[...

after:

{
  "runs" : [ ... ],
  "schemaVersion" : "1.0.0"
}

Consumers should read schemaVersion first and follow the version policy in the contract document.

Choosing the result reader

xcresulttool's legacy API — the way every version before 4.0 read result bundles — is deprecated and will be removed from Xcode. xchtmlreport now has two readers and picks one per run:

$ xchtmlreport --result-reader auto TestResults.xcresult    # the default

• auto prefers legacy while the toolchain still offers the legacy commands, and falls back to modern once they are gone (or when the probe cannot tell).
• modern forces the new-format reader on any toolchain.
• legacy forces the legacy reader; if the toolchain no longer provides the legacy commands this is an error, never a silent substitution.

The XCHR_RESULT_READER environment variable sets the default when the flag is absent — useful for forcing a whole CI job onto one reader.

Reports from the two readers are held byte-identical by a differential test suite, up to a short declared list of differences the new format cannot avoid:

• Attachment display names. The new format does not expose the user-supplied XCTAttachment name, so the modern reader labels attachments by their type (Screenshot, Video, File).
• Failure title prefixes. Legacy renders Assertion Failure at File.swift:12: message; modern renders File.swift:12: message — the new format pre-joins the string and drops the issue type.
• Wrapper groups. Legacy nests two extra levels (All tests / Selected tests, then <target>.xctest) that the new format does not have; the modern reader renders the natural flat tree.
• Group durations. The new format reports no duration for test suites and bundles, so the modern reader shows (0.00s) where legacy shows a real value.

--json output additionally carries testCase.arguments (Swift Testing @Test(arguments:) values), which only the modern reader can populate — the legacy format has no counterpart, so it is always [] there. See the schema contract for how consumers should compare reports across readers.

Exit codes

Starting in 3.0, xchtmlreport exits non-zero when part of a result bundle cannot be turned into a report. Earlier versions exited 0 and printed a success message even when parts of the report were missing, so pipelines had no way to detect an incomplete report.

Exit 0 means no faults were detected, which is not yet the same as a guaranteed complete report: some XCResultKit decode failures are not surfaced as faults today and are still only visible as messages on stderr. See #378 and the follow-on work it tracks.

Exit 1 covers failures to write the output. A missing -o directory is not one of them: it is created for you, intermediate components included, before any parsing or rendering starts. A directory that cannot be created — no write permission on the parent, or a file in the way — is an argument error (64), reported up front and naming the path.

The report is still written when faults occur. To restore the pre-3.0 behaviour and always exit 0 on faults, pass --lenient. --lenient does not affect exit 1 or 64.

Fastlane Support

https://github.com/TitouanVanBelle/fastlane-plugin-xchtmlreport

Contribution

Please create an issue whenever you find an issue or think a feature could be a good addition to XCTestHTMLReport. Always make sure to follow the Contributing Guidelines. Feel free to take a shot at these issues.

Special Thanks

Thank you to the original author of this tool, TitouanVanBelle! 🥳🎉

License

XCTestHTMLReport is available under the MIT license.