Lighthouse for the AI Age - A free, open-source website auditing tool that evaluates and optimizes websites for AI agent interactions.
With 1 billion+ monthly AI crawler requests (GPTBot, ClaudeBot, PerplexityBot), optimizing for AI visibility is no longer optional. SoberAI is the first specialized tool for ensuring your website works perfectly with AI agents, LLMs, and autonomous browsing systems.
| Category | Weight | Description |
|---|---|---|
| SSR Readiness | 20% | Server-side rendering for AI agents that don't execute JavaScript |
| Schema Coverage | 20% | Structured data (Schema.org) for AI understanding |
| Semantic Structure | 20% | HTML5 semantic elements and heading hierarchy |
| Content Extractability | 20% | Text readability and accessibility for LLMs |
| Machine Readability | 20% | AI crawler access (robots.txt, llms.txt, sitemap, OpenGraph) |
AI-powered recommendations default to an OpenAI-compatible key (OpenAI, NVIDIA NIM, or any /chat/completions endpoint, including bring-your-own-key mode that never stores the key server-side). Ollama local/cloud and Anthropic stay available as opt-ins. Without any key, audits complete with finding-based recommendations.
Download the latest release for your platform from the Releases page.
# Clone and install
git clone https://github.com/nitishagar/sober-ai.git
cd sober-ai
npm install
# Build frontend
cd frontend && npm install && npm run build && cd ..
# Set up database
npx prisma generate --schema=src/db/schema.prisma
npx prisma migrate deploy --schema=src/db/schema.prisma
# Start the server
node src/api/server.js
# Open http://localhost:3000/appnpm run electron:devStart the backend (Option 2), then load extension/ unpacked at
chrome://extensions (Developer mode). Audit the active tab from the
popup; keys stay in chrome.storage.local. See
extension/README.md.
Set an OpenAI-compatible key — environment, Settings page, or bring-your-own-key mode:
export OPENAI_API_KEY=sk-...Or go fully local: install Ollama, run
ollama pull qwen3:4b, and select ollama_local in Settings.
- Navigate to the Audit page
- Enter a URL (e.g.,
https://vercel.com) - Click Run Audit and watch real-time progress (auto-reconnects on network interruption)
- Review scores across 5 categories, findings, and AI recommendations
- Compare reports over time to track improvements
- Configure your LLM provider in Settings
Electron Shell
├── React Frontend (BrowserWindow)
├── Express.js Backend (in-process)
├── SQLite Database (Prisma ORM)
├── Playwright (browser automation)
└── LLM Provider (OpenAI-compatible default; Ollama local/cloud or Anthropic opt-in)
- Desktop: Electron
- Backend: Express.js + Node.js 22+
- Frontend: React + Vite
- Database: SQLite via Prisma
- Browser: Playwright
- LLM: OpenAI-compatible default (OpenAI, NVIDIA NIM, self-hosted gateways); Ollama or Anthropic opt-in
- Extension: Chrome MV3 tab audit (no build step)
npm test # Run all tests (Jest unit + integration suites)
npm run test:unit # Unit tests only
npm run test:integration # Integration tests only
npm run test:ci # CI mode (all tests, force exit)
npm run verify # Run tests + lint
npm run test:e2e # Playwright E2E tests (browser)
npm run verify:full # Unit + integration + lint + E2EFirst-time E2E setup (installs Chromium, ~400MB):
npm run test:e2e:installThe test suite covers:
- Unit tests: All 5 audit classes (score boundaries, severity, findings), scorer, validator, LLM analyzer, provider factory
- Integration tests: SSE streaming (reconnect, error paths), audit pipeline, settings, reports (CRUD + compare), provider routing, health
- E2E tests: Audit happy-path (real SSE + DB), reports list/detail, health banner, SSE reconnect branch (
tests/e2e/)
Full documentation available at nitishagar.github.io/sober-ai.
We welcome contributions! See the Contributing Guide for development setup and guidelines.
MIT License - see LICENSE for details.
Copyright 2025-2026 Nitish Agarwal
Version 0.4.0 | Changelog
