v0.15.7 - report footer placement

Moves the report's disclaimer footer to the true end of the document (after the per-page appendix).

v0.15.6 - Full report (PDF) export

The Print/PDF export is now a comprehensive, story-driven client report instead of just the score + action plan. Structure: cover + overall score, 01 Overview, 02 What's already working (your strengths), 03 What's holding you back (every finding, grouped + severity-rated), 04 What to fix, in order (+ a 90-day roadmap), 05 The result you can expect (current -> projected, per engine), and an appendix of every page. Light, editorial, print-optimised, and white-label aware. Button renamed 'Full report (PDF)'. No scoring changes.

v0.15.5 - cross-engine AI answer panel

New tool: analyze_ai_panel - aggregate several AI engines' answers to the same question into one cross-engine view. Shows where your brand is NAMED vs merely CITED (the 'cited != recommended' gap), share of voice, the competitor leaderboard, and your rank. Paste-based, no automation or keys. Also fixes a brand-detection false positive when the brand name is part of its own domain. No scoring changes.

v0.15.4 - anonymous usage analytics

The app now records ANONYMOUS, content-free usage so we can see how CITED Score is actually used and improve it. Default-on, disclosed, one-click opt-out (checkbox on the main screen).

Recorded (buckets / categories / booleans only): page-count band, site type, score band, whether a run was a re-run, feature use, launch + version.
Never sent: your URLs, domains, the sites you audit, page content, or anything identifying you. Re-runs are detected on your own machine - only a true/false leaves.

No scoring changes. Existing installs will show an update banner.

v0.15.3 - citation-gap fixes

Two fixes to the citation-gap tooling (MCP + engine). No scoring changes; existing installs will show an update banner.

• correlate: empty gaps fixed. Pages absent from a citations export were stored as None, but the gap filters test == 0, so opportunities (well-scored pages earning zero citations) was always empty for a normal Bing/Clarity export. Absent-from-export now counts as 0 reported citations, so the real opportunities surface.
• correlate + citation_coverage now crawl the whole site by default. A low page cap dropped deep blog/content pages (where most citations land), producing false gaps. These two data-comparison tools now default to full coverage (max_pages=0); pass max_pages only to cap a large site for speed.

v0.15.2 - Connect/Disconnect toggle + Recent reports capped at 5

• Connect / Disconnect toggle for Claude: the MCP card now lets you disconnect (removes only cited-score from Claude's config, keeps your other servers).
• Recent reports capped to the 5 most recent - older reports are cleared automatically; the Score-over-time trend data is preserved.
• Includes everything from v0.15.x.

Unsigned build: one-time Windows 'unknown publisher' warning on first run - More info -> Run anyway.

v0.15.1 - context-aware CSV export

The Export CSV button now exports the data of the tab you're viewing - Pages, Action Plan, Broken links, or Sitemap issues - and its label updates to match (e.g. 'Export broken links (CSV)'). Print / PDF still exports the whole report. Includes everything from v0.15.0.

Unsigned build: one-time Windows 'unknown publisher' warning on first run - More info -> Run anyway.

v0.15.0 - Sitemap Health, Site Map & trend monitoring

New in this release:

• Sitemap Health (Site structure tab): indexable pages MISSING from your sitemap, orphan pages with no internal links, and noindexed URLs that shouldn't be listed - AI can't cite a page it can't reach.
• Site Map: a visual link map (rings = crawl depth, colour = score, red ring = orphan).
• Score over time: a sparkline of your AI-citability across re-crawls.
• Scheduled re-crawls: a 'Schedule weekly re-crawl' button registers a Windows task so the trend fills in automatically.
• Everything from v0.14.x (MCP server + Connect to Claude, check_draft, AI-bot log monitor, reasoning tools).

Unsigned build: one-time Windows 'unknown publisher' warning on first run - More info -> Run anyway.

v0.14.1 - site-type classifier fix

Patch on v0.14.0.

• Fix: an agency/consultancy with a large content blog (many articles + service pages) is no longer mislabeled as blog/publisher - it now classifies as General (the neutral scoring profile). Real publishers still classify as Blog.
• Includes everything from v0.14.0 (MCP server + Connect-to-Claude, check_draft, log monitor, and the new reasoning tools).

Unsigned build: one-time Windows 'unknown publisher' warning on first run - More info -> Run anyway.

v0.14.0 - Use CITED Score inside Claude (MCP)

New: use CITED Score inside Claude, Cursor, or Claude Code.

• Connect to Claude Desktop in one click (in the app) - then just ask your AI to audit a site, check whether a draft will get cited, and more. 11 tools + guided prompts.
• check_draft - lint a draft's AI-citability before you publish.
• AI-bot log monitor (--monitor) - see which AI crawlers hit your site over time, from your server log.
• New reasoning tools: click-resilience, correlate, AI-influence estimate, agent-transaction readiness, AI-answer analysis.
• Opt-in anonymous usage sharing (off by default; tool counts only, never the sites you audit).
• Robustness + stress-test fixes.

Unsigned build: on first run Windows may show a one-time 'unknown publisher' warning - click More info -> Run anyway.