docs(research): add Research Edition page for researchers

[TimeToBuildBob]

Summary

Adds src/research/research-edition.rst, a researcher-facing page for the Research Edition, at the top of the existing Research section toctree (the section, participant instructions and Android instructions already exist). Also fixes two research/index.rst bullets that overstated things: "Category-only collection" (app names are kept in the current build) and "plus Android" (Android has no filter or release yet).

Final URL once deployed: https://docs.activitywatch.net/en/latest/research/research-edition.html

Companion blog post: ActivityWatch/activitywatch.github.io#73. Merge this PR first, then the post, so its docs link does not 404.

Verified

• Filter behaviour (longest-match category map, URL first then title, excluded, titles/URLs dropped before storage, non-browser titles dropped, optional research_app_category_map, Excluded for unmapped apps): aw_watcher_window/research_filter.py and config.py.
• App names retained in the current build (release build no longer injects the app map): brain notes on the app-name data contract and the b5 smoke test.
• Hostname rewritten to research-participant on export; export refuses unfiltered data: existing participant instructions and b5 hostname proof.
• Port 5667, separate data folder, badge: b5 smoke test and existing participant instructions.
• v0.14.0b5-research is the latest desktop research prerelease; Android research flavor (net.activitywatch.android.research, 5667): aw-android/mobile/build.gradle. No -research Android release exists.
• aw-import-screentime requirements: its README.
• Builds cleanly in a minimal Sphinx setup (the full site build was not run locally: missing extensions).

Please double-check (Erik)

1. Contact address [email protected] is unverified: it is the email on your public GitHub profile.
2. "iPhone and iPad: Planned" and "the research filter does not apply to that data path yet" come from internal task notes, not from the importer's code.
3. "A variant is mostly a category map, a tagged build, and a participant guide" is my synthesis of the pipeline, not documented elsewhere.
4. The existing Android participant instructions describe a sideload install, but no Android -research release is published; my page says so. Confirm that is the intended wording.
5. I did not add data-size estimates (brain notes have an unmeasured 1-5 MB per participant-week figure).

[greptile-apps]

Confidence Score: 5/5

[Low risk] Adds documentation page for research study edition.

The PR appears safe to merge, though the export description should distinguish non-browser events from categorized browser events.

Findings

1.  Non-browser categories overstated ▶

Summary

The PR adds a researcher-facing Research Edition page, links it from the Research section, and corrects the section’s platform and collection claims. The revised data-size section now identifies timestamps, durations, and retained application names, but overstates which window events have a study category.

_{Reviews (5) · Last reviewed commit: "docs(research): describe export as filte..."}

[TimeToBuildBob]

@greptileai review

[TimeToBuildBob]

🤖 AI code review

Safe to merge — no P0/P1 findings

Confidence 5/5

✅ No findings. The diff looks correct to me on this pass.

Previous review passes

commit	score	findings	engine	when
11641d2e33c9	5/5	0	llm	2026-09-29 18:59 UTC
0e86df5c730d	5/5	0	llm	2026-09-29 19:05 UTC

_{Reviewed dfdc188347ae · openrouter/deepseek/deepseek-v4-flash-0731 · llm (in-band) engine · 19s · about this reviewer}

_{Maintainer commands}

_{@TimeToBuildBob review (own line) — fresh review · @TimeToBuildBob fix — a worker acts on the findings. Once per comment; 👀 = received.}

[TimeToBuildBob]

@greptileai review

[TimeToBuildBob]

@greptileai review

[TimeToBuildBob]

@greptileai review

[greptile-apps]

Non-browser categories overstated. The default export also includes non-browser window events. As the page explains above, those retain application names without a study category unless the optional app-category map is configured. Saying that window events keep both an application name and a study category could lead researchers to expect a category for every window event.

Suggested change

	has a timestamp and duration; window events keep the application name and the
	study category, with titles and URLs already dropped. Extra watchers connected
	has a timestamp and duration; window events keep the application name by
	default. Browser window events also keep the study category, with titles and
	URLs already dropped. Extra watchers connected

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!