Vireloop - AI SEO Content Autopilot

What is Vireloop - AI SEO Content Autopilot?

Vireloop is an AI SEO content autopilot I built to help agencies run content for many client sites from a single engine. It researches and scores keywords, builds a 30-day content plan, generates publish-ready articles through an 8-step AI pipeline, sends drafts to a human review queue, and publishes across WordPress, Next.js, and Git-based static blogs, with per-organization security, analytics, and automated maintenance.

Technologies & Tools Used

Technical Highlights

• Next.js 16 - App Router with Server Components, Server Actions & background tasks (after()), React Compiler enabled
• React 19 - Component-driven UI, State-managed dashboards
• TypeScript - Strict typing, Scalable, typed data models
• TailwindCSS - Utility-first styling, Token-based light/dark design system
• Supabase - Postgres database, Authentication, Row-Level Security (multi-tenancy), Vault secret storage
• AI Integration - Anthropic Claude & Google Gemini, Multi-step content pipeline, Automated quality scoring
• Publisher Connectors - WordPress & WordPress.com (OAuth), Next.js / Supabase & REST webhook, Git static blog via GitHub API
• Zod - Runtime validation, Typed external boundaries
• Analytics - Google Search Console, Google Analytics 4, Content-decay tracking
• recharts - Interactive charts, Traffic & ranking visualizations
• sharp - Server-side image processing, Title overlays on featured images
• lucide-react - Consistent icon system, Clean UI visuals
• pg_cron & Edge Functions - Scheduled daily pipelines, Background workers

Key Features

• Architected a multi-tenant SEO automation platform (organization-scoped from day one) using Next.js 16, React 19, Supabase, and Tailwind CSS.
• Built a resumable, database-backed AI content pipeline: SERP analysis, brief, outline, draft, on-page SEO pass, image generation, and automated quality scoring with per-step persistence and retries.
• Designed a platform-agnostic Publisher abstraction with connectors for WordPress, WordPress.com (OAuth), Next.js/Supabase, REST webhook, and Git-backed static blogs.
• Developed a GitHub connector that commits Markdown articles to a repository via the GitHub API, triggering static-site rebuilds, enabling automated publishing to file-based blogs with no database.
• Integrated multiple LLM providers (Anthropic Claude and Google Gemini) behind clean, Zod-validated interfaces with cost and token logging.
• Implemented keyword research and scoring (value vs. difficulty), 30-day content planning, topic clusters with automatic pillar/cluster interlinking, and a drag-to-schedule content calendar.
• Integrated Google Search Console and Google Analytics 4 (service-account JWT) for real traffic, ranking, and content-decay insights.
• Secured all credentials with Supabase Vault and Row-Level Security, keeping service-role keys and secrets server-side only, with a separate super-admin control plane.
• Streamlined site onboarding with background website crawling and AI business-profile generation, plus a progressive 'finish setup' checklist that inherits organization defaults.
• Delivered a responsive, accessible dashboard with light/dark theming, real-time pipeline status, and complete loading, empty, and error states following a token-based design system.

Project Category