I built SoberAI because I kept seeing the same problem: AI crawlers (GPTBot, ClaudeBot, PerplexityBot) don't execute JavaScript. Your React SPA is literally a blank page to them. There was no tool to audit for this — so I made one.
SoberAI is an Electron app that scores your site 0-100 across four categories: SSR readiness, Schema.org coverage, semantic HTML structure, and content extractability. It renders your page twice with Playwright (JS on vs off) and diffs the content. LLM-powered recommendations are optional (Ollama local or OpenAI).
One command: npm install && npm run electron:dev
No accounts. No telemetry. MIT licensed. ~3K lines of JS.
Docs: https://nitishagar.github.io/sober-ai
Would love feedback on the scoring methodology and what auditors to add next.
I wrote about "Agentic Debt" and the new OWASP LLM07:2025 classification. The industry is finally converging on a standard: Prompts are not secrets or config.
Building in AI? Decide if you're shipping the "picks & shovels" (MCP server platform) or the "gold" (AI Agent) – your strategy depends on it! #AI #ProductStrategy #MCP #AIAgents #TechInnovation