I split the llms corpus into smaller files, added ai.txt, and opened a public /api/knowledge/summary endpoint so agents can pull canonical site data without scraping everything.
Pages should load quicker on repeat visits, WhatsApp and Instagram previews should work again, and iOS visitors get a gentle Add to Home Screen nudge after a few visits—not on first load.
Added
- Tiered llms files (stories, newsroom, full corpus) plus hi/ar/es indexes and ai.txt
- GET /api/knowledge/summary — public JSON for AI agents
- Linked JSON-LD (@id URIs) across Person, Article, FAQ, Speakable, ProfilePage, HowTo
- Pricing FAQ + Speakable + HowTo schema; hidden pricing summary for voice/search
- IndexNow ping on deploy (when INDEXNOW_API_KEY is set)
- Essay: Minimum Viable Movement
- iOS Add to Home Screen hint (after 5 sessions + engagement)
- AEO monitoring doc in project-documentation
Changed
- robots.txt lists llms/ai URLs; expanded AI crawler allowlist
- Slimmed llms-full.txt (~6 KB index); bodies moved to tiered files
- OG images normalised to 1200×630 PNG
- Article JSON-LD uses share description instead of subtitle
- Speed Insights v2; 1-hour ISR on pricing, about, stories
Fixed
- React hooks lint in iOS install hint
- Broken WhatsApp/Instagram previews (wrong OG size/description)
Work together?
View plans, ask a question, or book a call.