Verba - Automated Blogpost Agent

Verba is an AI-powered automated blog post generation platform designed for high-end, minimalist content creation. It leverages agentic workflows to research, write, and style blog posts according to sophisticated brand guidelines.

🏛️ Architecture

Verba is built with a modern, decoupled architecture designed for scale and intelligence.

Backend (FastAPI & LangGraph)

The core intelligence resides in the Python backend, structured as a series of specialized services:

• Agentic Orchestration: Uses LangGraph to manage a multi-node generation pipeline:
  • Context Loader: Injects brand guidelines into the agent's memory.
  • Research Node: Performs real-time web research via Tavily AI.
  • Planner Node: Crafts compelling, minimalist titles.
  • Writer Node: Generates structured Markdown content with research facts.
  • Image Nodes: orchestrates prompt generation and visual creation.
• LLM Layer: Integrated with OpenRouter, primarily utilizing Gemini 2.0 Flash for state-of-the-art text generation.
• Image Generation: Leverages Google Imagen 3 and Gemini 2.0 for premium header visuals, with deterministic Picsum fallbacks.
• Memory & Storage:
  • Brand Memory: Persistent session-based guidelines for voice, tone, and audience.
  • Storage Service: Thread-safe in-memory repository for posts and configuration, designed for easy migration to PostgreSQL.

Frontend (Next.js & Tailwind CSS)

A premium, minimalist user interface inspired by Aesop and Apple aesthetics:

• Next.js 15: Leveraging App Router and Server Components for optimal performance.
• Tailwind CSS: Custom design system focused on typography and negative space.
• HITL (Human-In-The-Loop): A dedicated workflow for reviewing, editing, and approving AI-generated drafts.
• Real-time Brand Context: A live editor to tune the AI's creative direction on the fly.

🛠️ Setup and Development

Backend

1. Navigate to backend/
2. Create and activate a virtual environment (python -m venv .venv)
3. Install dependencies: pip install -r requirements.txt
4. Set up .env file (copy from .env.example)
5. Run the server: python -m uvicorn app.main:app --reload

Frontend

1. Navigate to frontend/
2. Install dependencies: npm install
3. Set up .env.local file (copy from .env.example)
4. Run the development server: npm run dev

🚀 Future Features

Verba is evolving into a full-scale publishing powerhouse. Our roadmap includes:

• Persistence Layer: Migration to PostgreSQL and Redis for robust data handling and caching.
• CMS Integrations: One-click publishing to WordPress, Ghost, Medium, and Shopify.
• Social Media Engine: Automatic generation of tailored LinkedIn, Twitter, and Instagram snippets for every post.
• Multi-Agent Editorial Team: Specialized agents for:
  • SEO Agent: Optimizes content for search intent and keywords.
  • Fact Checker: Cross-references research facts for maximum accuracy.
  • Creative Director: Ensures visual and tonal consistency across all posts.
• Deep Analytics: Tracking engagement, read time, and performance metrics directly in the Verba Dashboard.
• Team Collaboration: Multi-user support with role-based access control (RBAC).

📄 License

This project is proprietary. All rights reserved.