feat(analytics): split og:image tracking by who fetched the image - #10475
Merged
Conversation
A social or messenger preview means a human shared a link. That is a product signal and belongs with the human numbers, so those events stay on anyplot.ai. A search or AI crawler fetching the same image is not a share. The distinction only starts mattering now. robots.txt on api.anyplot.ai was a blanket Disallow until #10472, so crawlers were never allowed to fetch /og/ at all — every og_image_view came from a link preview, and recording them all on the main site was right. Opening /og/ changes that: crawler fetches of 3,913 preview images would arrive in volume, drown the sharing signal, and inflate visitor counts on the site whose bot inflation audit 2026-07-08 already had to remove once. The split reuses detect_ai_agent rather than inventing a second rule: if it recognises the agent the fetch is machine-side and goes to bots.anyplot.ai, carrying the same assistant and kind props as bot_fetch so both events slice alike. Anything it does not recognise — Twitter, Facebook, Slack, WhatsApp — is a share and stays put. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Contributor
There was a problem hiding this comment.
Pull request overview
This PR refines server-side Plausible analytics so og_image_view events are recorded on different Plausible sites depending on who fetched the preview image: social/messenger link previews stay on anyplot.ai, while recognized AI/search crawlers are routed to bots.anyplot.ai with assistant/kind properties (reusing detect_ai_agent). This keeps human sharing signals from being diluted now that /og/ is crawlable.
Changes:
- Route
og_image_viewtoBOT_DOMAINwhendetect_ai_agent()recognizes the user agent; includeassistantandkindprops for machine-side events. - Add unit tests asserting the domain/props split for representative share-preview bots vs crawlers.
- Document the new split (and required Plausible property registration) and add a changelog entry.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| api/analytics.py | Splits og_image_view tracking between DOMAIN and BOT_DOMAIN using detect_ai_agent, adding assistant/kind props for machine-side fetches. |
| tests/unit/api/test_analytics.py | Adds async unit tests validating the new audience split and ensuring share-preview user agents remain on the main site without bot props. |
| docs/reference/plausible.md | Documents og_image_view audience split and notes additional property registration needed on bots.anyplot.ai. |
| CHANGELOG.md | Adds an [Unreleased] entry describing the new split behavior and rationale. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
…racking-by-audience # Conflicts: # CHANGELOG.md
Same convention fix as #10474, applied to this entry. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The change
anyplot.ai(unchanged)bots.anyplot.aiassistant,kindWhy now, and not before
A preview fetch means a human shared a link — a product signal that belongs with the human numbers. A crawler fetching the same image is not a share.
That distinction was academic until this release.
api.anyplot.ai/robots.txtwas a blanketDisallow: /until #10472, so crawlers were never permitted to fetch/og/at all: everyog_image_viewcame from a link preview, and recording them all on the main site was correct.Opening
/og/changes it. Crawler fetches of 3,913 preview images arrive in volume, and on the main site they would drown the sharing signal and inflate visitor counts — on the very site whose bot inflation an audit already had to remove once (2026-07-08, High #7, ~40% overcount).Implementation
The split reuses
detect_ai_agentrather than inventing a second classification rule. If it recognises the agent, the fetch is machine-side; machine-side events carry the sameassistantandkindprops asbot_fetch, so both events slice identically on the bot dashboard. Anything it does not recognise is treated as a share.Verification
pytest tests/unit— 1638 passed, 7 new: four preview agents asserted to stay on the main site withoutassistantprops, and three crawler classes (search / user-directed / training) asserted to land on the bot site with the right pairruff check+ruff format --check— cleanDashboard follow-up
Machine-side
og_image_viewevents carry the main site's image properties too, sopage,platform,spec,language,libraryand thefilter_*family need registering onbots.anyplot.aias well — noted indocs/reference/plausible.md. Without registration the events land but cannot be broken down.🤖 Generated with Claude Code