Bloggy 🤖

Not just a blog. A Knowledge Graph for the Agentic Web.
Built with Phoenix 1.7, LiveView, and Tailwind 3.

📖 The Story

The web is evolving. Content is no longer just for human eyes—it's efficiently consumed by AI agents, LLMs, and semantic search engines. Bloggy is the answer to this shift.

It is a modern publishing platform engineered from the ground up to serve two masters:

1. Humans: Who demand a beautiful, fast (LiveView-powered), and responsive reading experience.
2. Machines: Who require high-token-density, structured data ingestion (RAG-ready).

Stop building tailored APIs for every new AI tool. Bloggy speaks "LLM" natively.

📐 Architecture

Bloggy sits at the intersection of a traditional CMS and a Vector Data Source.

graph TD
    User[👩‍💻 Human Reader] -->|HTML/LiveView| LB[Load Balancer]
    Agent[🤖 AI Agent] -->|JSON/Context| LB
    
    subgraph "Bloggy Platform"
        LB --> Router
        
        subgraph "AI Layer"
            Router -->|Disovery| LLM(.txt map)
            Router -->|Ingestion| Context(RAG API)
        end
        
        subgraph "Core App"
            Router -->|Interactive| LiveView
            LiveView -->|CRUD| Schema
            Context -->|Read| Schema
        end
    end
    
    Schema --> DB[(Postgres)]

Loading

🚀 Key Features

For the AI Age

• 🗺 llms.txt Standard: A dedicated /llms.txt route provides a standardized content map, allowing AI crawlers to discover your content structure in milliseconds.
• 🧠 Context API: The /api/v1/posts/:id/context endpoint delivers your content stripped of HTML noise, enriched with metadata, and optimized for Vector Database embedding costs.
• 🔍 Semantic First: HTML structure is rigorously semantic, ensuring even standard scrapers get high-quality signals.

For the Production Engineer

• ⚡️ Phoenix LiveView: Developing rich, interactive UIs (like the admin dashboard) without writing a single line of React or client-side Javascript.
• 🔐 Built-in Auth: Role-based access control (RBAC) securely implemented with Guardian.
• 🎨 Tailwind CSS: A utility-first styling engine that scales.
• 📝 Markdown Core: Content is stored as pure Markdown, future-proofing your data against platform lock-in.
• 🐳 Docker Ready: Production-optimized Dockerfile included for instant deployment.

🛠 Getting Started

Get your instance running in less than 2 minutes.

Prerequisites

• Elixir 1.14+
• PostgreSQL 12+
• Node.js 18+

Quick Start

1. Install Dependencies

   mix deps.get
   cd assets && npm install && cd ..

2. Initialize Database

   # Designed to be idempotent
   mix ecto.setup

3. Launch

   mix phx.server

Visit http://localhost:4000 to see your new platform.

🛤 Roadmap

We are currently in Horizon 4 (Professionalization).

Phase	Focus	Status
Horizon 1	Foundation – Core CRUD, Database, Authentication.	✅ Complete
Horizon 2	Design – Tailwind implementation, Responsive UI.	✅ Complete
Horizon 3	AI Integration – llms.txt, Context API, RAG Hooks.	✅ Complete
Horizon 4	Scale & Ops – CI/CD, Documentation, Containerization.	✅ Complete
Horizon 5	The Future – Vector Search, Auto-Summarization.	🔄 In Progress

See full details in ROADMAP.md.

🧪 Testing & Quality

We maintain a rigorous standard of code quality to ensure reliability for both human and machine consumers.

# Run the full suite (Unit + Integration)
mix test

# Check code style
mix format --check-formatted

📚 Documentation

• ARCHITECTURE.md: Deep dive into the system boundaries and data flow.
• Context API Schema: Implementation details of the AI endpoints.

License

MIT © 2026 Bloggy Team.