Changelog

What shipped, and when.

Every material change to Aelo. Bug fixes, features, infra work. No PR-speak.

2026-07-07·v0.10.0feature

Forum Hub becomes strategic

  • Citation Map: new default view under Forum Hub that ranks the sources actually cited across your workspace scans. The Sage improvement over generic 'go post on Reddit' advice.
  • Discover source list expanded from 5 → 16 (added G2, Capterra, TrustRadius, Trustpilot, Medium, DEV, Substack, Product Hunt, AlternativeTo, GitHub, and more).
  • Per-source strategy notes, 'What to do' actionable Sage layer explaining how each source is weighted by LLMs.
  • Jump-to-Discover from any citation-map row closes the insight → action loop.
2026-07-06·v0.9.0feature

Editorial & distribution

  • Blog gains 'The 16 sources LLMs actually cite', flagship research post with ranked table + per-tier analysis.
  • RSS 2.0 feed at /rss.xml so Feedly + Reeder + newsletter tools can subscribe.
  • India Index prepped for August edition: 20 new brand workspaces across SaaS/Fintech/D2C/EdTech, allowlisted for public receipts.
2026-07-05·v0.8.0polish

SEO + Sage-completion pass

  • Sitemap 3 → 25 URLs; robots.txt env-driven; Sage OG image (1200×630) auto-served at /opengraph-image.
  • JSON-LD structured data emitted across marketing pages (Organization, WebSite, Article, SoftwareApplication, Dataset, Breadcrumb).
  • All product emojis stripped, Sage archetype rule enforced. Analytics no longer fabricates 'vs last week' deltas without baseline. /docs rewritten with real quickstart + install + event API + scan API + FAQ.
2026-07-05·v0.7.0feature

The trust layer, receipts everywhere

  • 'See the receipt' drawer wired into dashboard tiles + LLM Tracker platform cards + India Index brand rows, every metric one click from raw scans.
  • Public /methodology page publishes every formula (mention rate, avg position, health score, share of voice, verdict tiers) with strict thresholds, no smoothing.
  • Public India Index receipts: anyone can click any ranked brand to see the exact prompts + verbatim LLM responses. No auth required.
2026-07-05·v0.6.0feature

India AI Visibility Index

  • New /india-index, Preview Edition with 6 brands across SaaS/Fintech/D2C/EdTech, ranked by real Gemini scans. Every number auditable.
  • Public /api/india-index endpoint returns the current edition; /api/india-index/scans?brand=X serves the receipts.
  • Zoho DOMINANT, Byju's INVISIBLE, real, honest, publishable.
2026-07-05·v0.5.0feature

Full marketing site + install snippet

  • 17 new marketing pages shipped: landing, product, pricing, solutions (founders / marketing / agencies / india), manifesto, about, contact, docs, blog, changelog, customers, privacy, terms, security.
  • Contact form wired to Resend at /api/contact (delivery gated on RESEND_API_KEY).
  • New /dashboard/settings?tab=install page with copy-pasteable snippet + framework-specific placement guide.
2026-07-04·v0.4.0feature

The receipt is live

  • Interventions model shipped, action + baseline + follow-up + verdict, per intervention.
  • New /dashboard/interventions page with per-row receipts and Measure Impact button.
  • Forum Hub → intervention loop-closer: marking a thread as posted auto-creates an intervention with a baseline snapshot.
2026-07-04·v0.3.0feature

Design system + brand archetype

  • New Sage + Magician design system: beacon-ivory accent, tabular numerals default, spectrum-rule flourish reserved for receipts.
  • Refined mark and lowercase wordmark, applied across landing, dashboard, and auth.
  • PROVE nav group added, signalling the closed-loop philosophy.
2026-07-04·v0.2.0fix

Honest data policy enforced end-to-end

  • Killed the silent mock-data pathway: llm-scanner no longer fabricates a scan when providers fail.
  • Removed the mock → gemini masquerade at all three write sites.
  • Dashboard health-score key mismatch fixed, real numbers now flow through.
2026-07-03·v0.1.0infra

Migrations 015, 016, 017

  • 015, interventions table with baseline_snapshot / impact_snapshot / impact_summary.
  • 016, llm_scans extended with brand_variants, sentiment_score, sentiment_reason, list_items, confidence.
  • 017, public.users RLS infinite recursion (42P17) killed with a SECURITY DEFINER helper.