What kinds of HTML report can I publish with Comma?

Any of them — Comma renders the HTML you give it rather than parsing a specific format, so anything that writes a self-contained page or a folder with an index.html works. There are 146 step-by-step guides covering 11 categories: test results, coverage, security scans, code quality, notebooks & data, charts & maps, performance & bundles, accessibility, docs & api references, slides, sharing & hosting.

Does my tool need to be on the list?

No. The guides exist because each tool has its own quirks — where it writes output, which references break when the file is opened from disk. The publishing step is identical for a tool with no guide: upload the entry HTML, upload the folder alongside it, get a URL.

Why do so many HTML reports break when opened from the filesystem?

Because they load something as a separate request — a results JSON, a search index, a navigation fragment — and browsers block those requests from file:// origins. Publishing serves the whole bundle over HTTP with the references rewritten to the uploaded copies, which is why the interactive parts start working again.

Can reports be published automatically from CI?

Yes. PATCH the same report id from any pipeline and the suite keeps one stable URL with a revision per run. Routines can also re-run a report on a schedule without a pipeline at all.

Report types

Anything that writes HTML.

Comma renders the HTML you give it rather than parsing a format, so the list below is not a list of integrations — it is 146 guides to the specific quirks of 146 tools. Where each one writes its output, which references break when you open it from disk, and what to upload alongside the entry file.

Test results

Unit, integration and end-to-end runs. Publish red builds too — those are the ones with a reader.

Coverage

Line and branch coverage trees, where the drill-down is the whole point.

Security scans

Dependency advisories, SAST findings and infrastructure scans. Keep these on a restricted access setting.

Code quality

Linters and static analysis, where a backlog only gets cleared if people can claim it.

Notebooks & data

Analysis that has to stay interactive to be worth reading.

Charts & maps

Single-file visualizations, and how to keep them from going blank in transit.

Performance & bundles

Load tests, profiles and bundle treemaps — regressions are visible as revisions.

Accessibility

The findings split across design and engineering, so the report has to reach both.

Docs & API references

Generated documentation and static sites — most of which break when opened from the filesystem.

Slides

Decks that lose their point the moment they are flattened to PDF.

Your tool isn’t listed

It still works. The guides differ because each tool writes its output somewhere different and breaks in its own way; the publishing step is the same one every time — upload the entry HTML, upload the folder alongside it, get a URL. Start with sharing an HTML report or publishing from CI.