- AI
- Frontend
- Automation
- Backend
- Fullstack
klarava – AI understandability check for websites
Image credit: Majd Edriss
Challenge
How well do AI systems understand company websites? With klarava, I explore technical accessibility, content clarity and AI interpretation. Findings should be traceable to the pages that were checked.
Result
klarava has a publicly usable version and remains in development. The free check creates a browser report without registration or an email address, with evidence from the checked pages and prioritised next steps. A higher score does not guarantee mentions in AI answers.
Solution
Three separate analysis layers combine technical checks, rule-based content analysis and an AI interpretation with source evidence. The report identifies affected pages and includes actions and drafts for an AI profile, FAQ and optional llms.txt. These drafts need professional review. The project articles document earlier development stages, including the API extension.
Technologies
- Node.js
- OpenAI gpt-4o-mini
- Supabase (EU)
- Vanilla JS
- REST API (OpenAPI)
- RapidAPI
- VPS (systemd + Caddy)
- nodemailer
- Cloudflare Turnstile