Changelog
Everything Seeklyn has shipped, newest first. One indie developer, building in public.
AI Search Intelligence suite
- The dashboard is regrouped by job: Setup → Observe → Measure → Compete → Diagnose → Impact, with one-off utilities under Tools and formulas under How it works
- Observe: Answers shows what each AI engine actually said for your prompts, with the sources it cited
- Measure: AI Visibility (mention, citation and recommended rates, stability) and Markets (country by country)
- Compete: Competitors (citation share of voice, prompt gaps) and Sources (the sites and pages AI cites instead of you)
- Diagnose: Opportunities ranked by impact, Narrative & Accuracy (how AI describes you, checked against your declared facts), Content Map and Site Health (crawl issues, AI-crawler access, Core Web Vitals)
- Impact: Search Performance (Google rankings for your prompts), Backlinks and Business Impact — third-party data stays labelled simulated until the integration is connected
- Every number traces to its evidence (metric → prompt → AI answer → cited URL), shows its raw count, and carries a label: your data, sample data, or simulated
- No proprietary 0–100 AI score anywhere; formulas are one click away and on the Methodology page, with CSV / JSON export
- Engines & Models lists the AI engines your prompts run on — ChatGPT, Google AI Overviews, Perplexity and more
- Old dashboard links (Brand Visibility, Compete, Prompts, Citations) redirect to the page that now owns that job
Landing page redesign
- Hero shows the Overview command center: every number opens its evidence, fixes ranked by impact
- New AI Search Intelligence section walks the five jobs — Observe, Measure, Compete, Diagnose, Impact
- The content loop and "the shift" are rebuilt as animated stories on the new dashboard
- New editor showcase: paste a formatted draft, watch every section get scored, apply a fix and see the score climb
- "Why watch a number you can’t act on?" is now a click-the-drop demo: a dashboard dead end vs a Seeklyn work order
- How it works rebuilt: an interactive dashboard tour, the evidence chain, the loop step by step and a scoring lab
- Features, pricing, the capability matrix, FAQ and the about page updated to describe the suite
Calmer dashboard loading
- When a report is slow to load you now see a soft glass placeholder shaped like the report, never a technical error
- Reports retry quietly in the background and fade in the moment they arrive
- If loading takes longer than usual, a single “Try again” appears — save and delete actions explain themselves in plain words
Growth Phase 3 — trust surfaces
- Live honest proof counter under the hero: shows total headline + citation checks only when the real total is meaningful (never a fake number)
- Public stats API endpoint with a 5-minute in-process cache (no database hammering)
- Founder strip between the stat band and pricing — the indie story made visible
- 30-day money-back guarantee wired into the refund policy page
Conversion & UX fixes
- Hero now doubles as the analyzer — a no-JS GET form feeds the ?title= deep link directly into Headline Studio
- Hero subhead rewritten with price anchor, guarantee, and honest admission
- Upgrade CTA links to /dashboard/plan with honest copy (no longer dead-ends)
- Cited quota shown up-front; upgrade wall no longer hides the last result
- In-app "Request Pro" flow replaces the mailto dead-end
- Pricing "Get Pro" button routes to the plan page, not the free analyzer
- ProTip "Learn how" links to the power-words glossary page
- Real share action + honest celebration threshold on the score banner
Brand Visibility hub + dashboard polish
- Unified Brand Visibility hub with Leaderboard, Prompts, and Sources tabs
- Share-of-Voice trend chart, competitor drill-down, and contextual diff
- Projects hub redesign — stop minting empty guest accounts
- Login rebinds to canonical account by email; logout wired correctly
- Tour/profile modal token scoping, post-login redirect, landing CTAs fixed
Hornbill brand identity
- Hornbill visual identity launched across the site (the "is anyone home?" brand signal)
- Hornbill extended to the app, invite emails, PDF reports, badge, share card, OG image, and email templates
- Brand folklore rewritten to be location-free for an international audience
- In-product visibility journey added: Call → Rain → Abundance progression with tour mascot
Google sign-in
- Google OAuth 2.0 + PKCE sign-in wired up, reusing the existing cookie identity
Dashboard revamp — workspaces, team & billing
- Workspaces schema + store with personal-workspace backfill (migration 0018)
- Hash-at-rest workspace invites store
- Workspace-visible projects, viewer read-only, owner-plan quotas
- Team page with invite acceptance flow (email-bound)
- Plan & Billing page with live usage meters
- Unique owner constraint + race-safe workspace ensure
Recommendations engine
- Recommendations schema + store (migration 0017)
- Rule engine core with citation and citation-change rules
- Ops rules and signal gathering, cron generation
- Recommendations list + status transitions API
- Recommendations page in the dashboard with status workflow + fix deep-links
Dashboard revamp — brand ranking & competitors
- Competitors schema + store CRUD, suggestions, recategorize (migration 0016)
- Brand ranking query with tone bands (positive / neutral / negative mention scoring)
- Write-time tone bands with evidence spans on all mentions
- Competitors management page and brand ranking on overview
- Shared AFINN valence package extracted from the headline scorer
Dashboard revamp — citations, deltas & sources
- Citations API: deltas, sources, URL detail, and per-prompt citation refs
- Citations page with delta tabs and URL detail drawer
- Overview citation sections and prompt citations tab
- Prompt deep-link, backfill rank-collision fix, tone dedup fix