I just built a fully offline, streaming-enabled AI content generation system running entirely on Ollama and wrapped in a gorgeous GhostOS UI. No API keys, no cloud dependencies – just raw, sovereign power from your machine.
⚙️ What It Is
GhostOS ContentAI is an AI-powered content generation platform that:
- Runs 100% locally with Ollama
- Supports multiple content categories (Text, Ads, Visuals, Docs)
- Offers streaming output with true LLM tokens
- Includes a persona selector (Ghost King, Academic, Corporate, etc.)
- Feels like a real operating system interface
💻 Built With
- 🧠
mannix/llama3.1-8b-abliteratedvia Ollama - 🛠️ Next.js 15 (App Router)
- 🎨 Tailwind CSS v4
- 💬 React + Zustand
- 🖼️ Lucide + Framer Motion for polish
🎭 Persona-Based Prompt Control
The Flame doesn’t just generate content. It channels it.
Choose from 8 distinct tone options:
- 👑 Ghost King (mystical, powerful, divine)
- 📚 Academic
- 💼 Corporate
- 🤡 Playful
- 🧑🔬 Technical
- 💬 Friendly
- 🎨 Creative
- ⚖️ Neutral
Each persona injects context and tone before generation.
🚀 Features
- ✅ Local-only (no cloud API)
- ⚡ Super fast response time
- 🔁 Real-time streaming output
- 🧠 Model switcher UI (supports multiple Ollama models)
- 🔐 Private, offline, and extendable
🧪 Use Cases
- Generate eBooks, reports, email copy, blog posts
- Build PDFs, proposals, contracts, brand guides
- Create ad scripts, banners, or tweet copy
- Auto-generate dev docs, product specs, or FAQs
📦 Repo Access
This project is part of a bigger system called the GodsIMiJ Empire – a sovereign AI ecosystem. I’ll be open-sourcing parts of it soon.
Drop a comment if you want:
- Early repo access
- A walkthrough of the architecture
- Your own persona-customized Flame build
🌐 Preview
Powered by GhostOS • Eye of Kai • Mirror of Code
Built for the sovereign coders, flame-born creatives, and rebels of the new digital age.
💬 Feedback?
I’m currently mentoring a small group of devs learning to integrate AI the right way. Want in? I’m offering:
- Free starter templates
- Private repo reveals
- Custom AGA GUIs (if the whisper calls)
5 slots max – DM or drop a comment if you're ready to learn the Way of the Flame.
Top comments (1)
sounds like a cool project 💪🏽 I would like access to git repo, to potentially steal some ideas
as I am also building offline first all in browser ai client, mine is meant for smaller tasks and is based on chrome built in gemini nano but users would be able to add api keys for a few most popular models if they need something with more compute capability
Some comments may only be visible to logged-in visitors. Sign in to view all comments.