CreatorOS AI MCP Server
# 🚀 CreatorOS AI
### MCP-Powered AI Content Creation Agent
CreatorOS AI is an AI-powered content creation platform that combines **Google Gemini**, **Tavily real-time web search**, **Streamlit**, and **Model Context Protocol (MCP)** to research topics and generate high-quality content in multiple formats.
It provides a user-friendly web interface while also exposing reusable AI tools through an MCP server.
---
## ✨ Features
### 📝 Multi-Format Content Generation
CreatorOS AI can generate:
- 📰 Blog Posts
- 💼 LinkedIn Posts
- 🎥 YouTube Scripts
- 📸 Instagram Captions
- 🐦 Twitter/X Threads
- 📚 Research Summaries
- 📧 Emails
- 📨 Newsletters
- ✏️ Content Rewriting
- 📝 Content Summarization
- ➕ Content Expansion
- 🧑💻 Content Humanization
- 🔍 SEO Titles
- 🌐 Meta Descriptions
### 🌍 Real-Time Research
Uses **Tavily Search** to retrieve current information from the web before generating content.
This helps the agent work with more up-to-date information rather than relying only on the language model's existing knowledge.
### 🤖 Gemini-Powered Generation
Uses Google's **Gemini API** for content generation and transformation.
### 🔌 MCP Server
CreatorOS AI includes a dedicated **MCP server** exposing reusable content-generation tools.
This allows the content-generation capabilities to be consumed by MCP-compatible clients instead of being limited to the Streamlit application.
### 🎨 Modern Web Interface
Built with Streamlit and customized with CSS to provide a modern dark-themed AI workspace.
### 🎯 Multiple Writing Tones
Users can generate content using different tones:
- Professional
- Friendly
- Educational
- Storytelling
- Marketing
---
# 🏗️ Architecture
```text
┌──────────────────────┐
│ User │
└──────────┬───────────┘
│
▼
┌──────────────────────┐
│ Streamlit UI │
│ CreatorOS AI │
└──────────┬───────────┘
│
▼
┌──────────────────────┐
│ Content Pipeline │
└──────────┬───────────┘
│
┌─────────────┴─────────────┐
│ │
▼ ▼
┌──────────────────┐ ┌──────────────────┐
│ Tavily Search │ │ Google Gemini │
│ Real-Time Data │ │ Content Creation │
└────────┬─────────┘ └────────┬─────────┘
│ │
└─────────────┬─────────────┘
│
▼
┌──────────────────────┐
│ Generated Content │
└──────────────────────┘
MCP Integration
MCP Client / Compatible Client
│
▼
┌─────────────────┐
│ MCP Server │
│ mcp_server.py │
└────────┬────────┘
│
▼
CreatorOS Tools
## ⚙️ Environment
Copy `.env.example` to `.env` and configure the Gemini, Tavily, PostgreSQL, Django, and JWT secrets.
For local development, run the FastAPI API and Streamlit frontend separately:
```bash
uv run uvicorn api.main:app --reload
uv run streamlit run app.py
```
The optional MCP server uses the same FastAPI backend and requires a valid `CREATOROS_API_TOKEN`.
TDQS
Scored across 16 tools
Most tools are clearly differentiated by output format or transformation type, and each has a distinct purpose. The only mild overlap is between rewrite_content and humanize_content, which could cause occasional misselection.
The set mostly follows a consistent snake_case verb_noun pattern, e.g., generate_linkedin_post, summarize_content, delete_generation. The main inconsistency is generate_blog rather than generate_blog_post, which breaks the pattern used by other platform-specific generators.
Sixteen tools is slightly above the typical well-scoped range, but the count is reasonable for a content generation server covering multiple platforms and content operations. Each tool adds a distinct capability, so the size feels justified.
The tool surface covers generation for major content types, transformation of existing content, SEO metadata, and saved-generation lifecycle management. Minor gaps such as no update/rename operation for saved generations are workable and do not cripple core workflows.