ADA Turbo
Official
README.md
# ADA Turbo — Agency OS & Pixel Office
<div align="center">
[](https://modelcontextprotocol.io/)
[](https://pypi.org/project/ada-turbo-ai-mcp/)
[](https://python.org)
[](LICENSE)
[](https://smithery.ai)
[](https://github.com/adacreativeco/ada-turbo-ai-mcp/stargazers)
[](https://adacreative.co/vaka-analizleri/ada-turbo-mcp)
<br/>
**🇺🇸 English Documentation | 🇹🇷 [Türkçe Dokümantasyon](README.tr.md)**
<p align="center">
<strong>Transform your IDE into a full-scale creative agency.</strong><br/>
ADA Turbo delivers the <strong>ADA Creative Co.</strong> agency operating system with 26 specialized AI roles via <strong>MCP (Model Context Protocol)</strong>, paired with a real-time retro CRT <strong>Pixel Office Visualizer</strong>.
</p>
</div>
---
## 🎬 Live Office & Animation Preview
<div align="center">
<img src="animasyonlar/ada-turbo-walk-preview.gif" alt="Pixel Office Characters Walking Animation" width="100%" />
</div>
<br/>
| 🏢 Pixel Office Visualizer | 🧪 CRT Workflow Playground |
| :---: | :---: |
|  |  |
---
## ⚡ Quick Start (Run in 10 Seconds)
### Option A: Zero-Install with `uvx` (Recommended)
You can run ADA Turbo instantly without cloning or managing virtual environments:
```bash
uvx ada-turbo-ai-mcp
```
*(Or launch just the web visualizer: `uvx ada-turbo-ai-mcp --web`)*
---
### Option B: Standard Git / Python
```bash
# 1. Clone the repository
git clone https://github.com/adacreativeco/ada-turbo-ai-mcp.git
cd ada-turbo-ai-mcp
# 2. Install dependencies
pip install -r requirements.txt
# 3. Start Unified Dual-Mode (MCP stdio + Web Visualizer)
python server.py
```
Open [http://localhost:8000](http://localhost:8000) (auto-fallback to `8001`, `8002`... if occupied) in your browser.
---
## 🚀 Key Features
- **Unified Dual-Mode Architecture:**
- **MCP Mode (Default):** Runs as an MCP stdio server. Concurrently starts the Pixel Office Web Server in the background. Seamlessly connects to Antigravity, Claude Desktop, Cursor, Claude Code, and Windsurf.
- **Pixel Office Web Mode:** Starts a local, highly-optimized retro CRT-effect web workspace (`python server.py --web`).
- **⚡ Real-Time Server-Sent Events (SSE):**
- Zero-latency event streaming (`/api/events`). When an agent is triggered in your IDE, the corresponding character visually walks to their desk in real-time!
- **🧠 Direct Live LLM Engine:**
- Multi-provider AI execution directly in the browser via `/api/llm-generate`:
- 🟢 **Built-in Template Engine (Offline):** Instant access to professional agency workflow templates without API keys.
- ✨ **Google Gemini:** `gemini-2.0-flash`, `gemini-1.5-flash`, `gemini-1.5-pro`
- 🧠 **OpenAI:** `gpt-4o-mini`, `gpt-4o`, `o3-mini`
- ⚡ **Anthropic Claude:** `claude-3-5-sonnet-20241022`, `claude-3-5-haiku-20241022`
- Safe local credential storage in browser `localStorage`.
- **Developer Tools (CRT Console Modals):**
- **Playground (Workflow & Live AI Tester):** Select commands and test tasks with offline templates or live LLM models with typewriter output.
- **AI Model Configurator:** Switch providers, enter API keys safely, and select model presets.
- **Setup Wizard:** Dynamically outputs ready-to-copy configuration blocks for Cursor, Antigravity, Claude Desktop, and Claude Code.
- **Pixel Characters & Animations:** 26 unique agency characters with idle (breathing, blinking) and walking cycles across custom office floors.
- **Full Bilingual Support (TR / EN):** One-click toggle for all UI elements, status badges, modals, agent prompts, and knowledge references.
---
## 🔌 IDE Integration (MCP)
### 1. Antigravity
Add to your `mcp_config.json`:
```json
{
"mcpServers": {
"ada-turbo": {
"command": "uvx",
"args": ["ada-turbo-ai-mcp"]
}
}
}
```
### 2. Cursor
Add to `~/.cursor/mcp.json` (global) or `.cursor/mcp.json` (project-level):
```json
{
"mcpServers": {
"ada-turbo": {
"command": "uvx",
"args": ["ada-turbo-ai-mcp"]
}
}
}
```
### 3. Claude Desktop
Add to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"ada-turbo": {
"command": "uvx",
"args": ["ada-turbo-ai-mcp"]
}
}
}
```
### 4. Claude Code (CLI)
```bash
claude mcp add ada-turbo -- uvx ada-turbo-ai-mcp
```
*(Note: If using local clone, replace `"command": "uvx"` and `"args": ["ada-turbo-ai-mcp"]` with `"command": "python"` and `"args": ["/ABSOLUTE/PATH/server.py"]`)*
---
## 👥 26 Agency Roles & Specializations
ADA Turbo organizes 26 specialized agents across 5 key departments:
1. **Strategy & Brand:** Brand Director, Brand Strategist, Consumer Insight Lead, Naming Specialist, Creative Technologist.
2. **Creative & Design:** Creative Director, Senior Copywriter, Art Director, UX/UI Lead, 3D/Motion Designer.
3. **Marketing & Growth:** Growth Marketing Director, Performance Marketing Specialist, SEO/Content Strategist, Social Media Lead, CRM & Retention Manager.
4. **Client & Operations:** Account Director, Senior Account Manager, Agency Producer, Operations Director, Traffic Manager.
5. **Analytics & Tech:** Chief Data Officer, Marketing Data Analyst, Full-Stack Lead, AI Solutions Architect, QA & Delivery Lead.
---
## 📂 Architecture
```
ada-turbo-mcp/
├── server.py ← Unified dual-mode entry point (MCP + Web)
├── index.html ← Retro CRT Pixel Office single-page UI
├── pyproject.toml ← Package metadata & entry points
├── smithery.yaml ← Smithery.ai 1-click install configuration
├── requirements.txt ← Project dependencies
├── references/ ← Bilingual domain knowledge bases (.md)
│ ├── strategy-brand.md / strateji-marka.md
│ ├── creative-team.md / yaratici-ekip.md
│ ├── marketing-growth.md / pazarlama-buyume.md
│ ├── client-operations.md / musteri-operasyon.md
│ └── analytics-product-tech.md / analitik-urun-teknik.md
├── karakterler/ ← Pixel character graphics and generator scripts
├── animasyonlar/ ← Character walk/idle spritesheets & GIF previews
├── office-bina/ & office-zon/ ← Procedural pixel art building generation assets
├── skill/ ← Pre-packaged .skill distribution bundle
└── src/ ← Python modules
├── mcp_server.py ← FastMCP server definitions & tool registrations
├── web_server.py ← Multi-threaded HTTP server, SSE broadcaster & LLM proxy
└── workflow_manager.py ← Command routing, action listener dispatch & templates
```
---
## 📄 License
Distributed under the PolyForm Noncommercial License 1.0.0. See [LICENSE](LICENSE) for details.
Developed with ❤️ by **[ADA Creative Co.](https://github.com/adacreativeco)**
TDQS
A3.5/5.0
Scored across 3 tools
Disambiguation5/5
Each tool has a clearly distinct purpose: listing commands, returning rules, and returning workflow instructions. No overlap or ambiguity.
Naming Consistency4/5
All tools use the 'ada_' prefix, but 'ada_workflow' mixes Turkish and English while the others are entirely Turkish. Mostly consistent with a minor deviation.
Tool Count4/5
With 3 tools, the set is small but seems appropriate for the narrow domain of ADA Turbo commands and workflows. No unnecessary tools.
Completeness4/5
The tools cover listing commands, returning rules, and workflow instructions. Minor gap: no tool to update or add new commands, but core functionality is present.
Maintenance
ActivitySlowing
ResponsivenessNo issues