Content Gardener AI (SEO Content Platform)

Web Dev
Full Stack
UI/UX

Tech Stack

Typescript
Next.js
React
Node.js
express.js
MongoDB
Tailwind CSS
Machine Learning
REST API

Description

Content Gardener AI is a full-stack platform for creating and managing SEO content with an AI assistant. The frontend is a Next.js protected portal with orgSlug and websiteSlug scoped URLs, while the backend provides the AI generation, scraping/indexing, chat notifications, WordPress publishing, and subscription/billing APIs.

The architecture uses Express with tsoa route registration (plus Swagger) and dependency injection (tsyringe) to keep long-running workflows orchestrated and data correctly scoped per website and organization. Users see progress through streaming updates using SSE, with generation stages rendered as they complete rather than waiting for a single blocking response.

A core advanced feature is multi-stage AI article generation using RAG. The system detects the request type (article vs ideas vs generic chat) and runs a structured pipeline (title checks → analysis → outline → section expansion → consolidation/enhancement → tags → save to the library) powered by OpenAI, with prompt-engineered templates at each stage, tracking token usage and persisting generation logs. For retrieval, it scrapes each website with Cheerio, chunks content, creates embeddings via OpenAI, and grounds outputs by retrieving relevant passages through MongoDB Atlas vectorSearch (with similarity scoring and fallback behavior when relevance is low).

For publishing, the platform integrates directly with WordPress. It stores WordPress connection configuration, syncs categories, generates AI category suggestions, and posts articles to WordPress using either WordPress.com OAuth or self-hosted REST credentials—then maps WordPress post status/IDs back into the app’s internal article workflow so the dashboard remains consistent after publishing.