Content
<div align="center">
# 🎓 LITERA-AI (Model Context Protocol Ecosystem)
### *Ekosistem kecerdasan buatan operasi pendidikan terintegrasi & sistem tutor adaptif pertama berbasis MCP (Model Context Protocol).*
[](https://python.org)
[](https://modelcontextprotocol.io)
[](https://huggingface.co/unsloth/DeepSeek-R1-Distill-Llama-70B-GGUF)
[](https://kurikulum.kemdikbud.go.id)
[](LICENSE)
<br/>
[📋 Overview](#-overview) • [✨ Fitur Utama](#-16-core-mcp-modules) • [🏗️ Arsitektur](#%EF%B8%8F-arsitektur-sistem) • [🚀 Cara Memulai](#-cara-memulai) • [🔌 Integrasi Client](#-integrasi-client-mcp) • [🤝 Kontribusi](#-lisensi--penulis)
---
</div>
## 📋 Overview
**LITERA-AI MCP** adalah platform server kecerdasan buatan terpadu berbasis **Model Context Protocol (MCP)** yang dirancang khusus untuk memodernisasi ekosistem pendidikan nasional dan internasional.
Dengan menghubungkan server MCP ini ke IDE/Client seperti **Claude Desktop**, **Antigravity IDE**, **Cursor**, atau **VS Code**, pengguna mendapatkan akses instan ke **16 Modul Spesialis MCP Pendidikan** dan **61+ Tools Otomatisasi Pembelajaran**.
> [!NOTE]
> **Fleksibilitas Mesin AI:** Memiliki dual-engine otomatis yang mendukung **Ollama Lokal** (*Offline & Private* via model `Litera-AI.gguf`) dan **OpenRouter Cloud API** (*High Performance*).
---
## 🏗️ Arsitektur Sistem
```mermaid
graph TD
subgraph ClientLayer ["🖥️ MCP Client Applications"]
Claude["Claude Desktop"]
IDE["Antigravity IDE"]
Cursor["Cursor / VS Code"]
end
subgraph ProtocolLayer ["⚡ Model Context Protocol (JSON-RPC 2.0 stdio)"]
MCPServer["LITERA-AI Master MCP Server <br/> (litera_mcp/server.py)"]
end
subgraph ModuleLayer ["🧩 16 Specialised MCP Educational Modules (61 Tools)"]
M1["1. Student Profile MCP"]
M2["2. Adaptive Learning MCP"]
M3["3. Assessment & Essay Rubric MCP"]
M4["4. Teacher & Modul Ajar MCP"]
M5["5. Parent & Weekly Report MCP"]
M6["6. AI Diagnostic (CNN + DKT) MCP"]
M7["7. Question Generator (HOTS/AKM/PISA) MCP"]
M8["8. Curriculum & Alignment MCP"]
M9["9. Socratic AI Tutor MCP"]
M10["10. Academic Research MCP"]
M11["11. Coding Education MCP"]
M12["12. Learning Analytics & Knowledge Graph MCP"]
M13["13. Content (Slides/Mindmap/Flashcards) MCP"]
M14["14. Voice & Pronunciation Scoring MCP"]
M15["15. School Administration MCP"]
M16["16. AI Gateway Engine MCP"]
end
subgraph AIEngineLayer ["🧠 Dual AI Execution Engine"]
Ollama["Ollama Engine Lokal <br/> (Model: Litera-AI.gguf / 70B)"]
OpenRouter["OpenRouter Cloud API <br/> (DeepSeek-R1 / Claude / GPT-4o)"]
end
ClientLayer -->|JSON-RPC 2.0 Stdio| ProtocolLayer
ProtocolLayer --> ModuleLayer
ModuleLayer --> AIEngineLayer
```
---
## ✨ 16 Core MCP Modules
| Icon | Modul MCP | Deskripsi & Kapabilitas Utama | Status |
| :---: | :--- | :--- | :---: |
| 🧑🎓 | **Student Profile MCP** | Pengelolaan profil murid, pemetaan gaya belajar (VARK), dan skor kemajuan. | `Active` |
| 📈 | **Adaptive Learning MCP** | Rekomendasi materi kustom & penyesuaian kesulitan dinamis (Dynamic Difficulty Adjustment). | `Active` |
| 📝 | **Assessment MCP** | Pembuatan ujian, evaluasi esai otomatis, rubrik analitik, dan tes plagiarism. | `Active` |
| 👨🏫 | **Teacher MCP** | Dashboard kelas, pembuat Modul Ajar (RPP Plus) Kurikulum Merdeka & Silabus. | `Active` |
| 👨👩👧 | **Parent MCP** | Laporan perkembangan anak mingguan/bulanan & panduan pendampingan orang tua. | `Active` |
| 🔬 | **AI Diagnostic MCP** | Analisis tingkat penguasaan konsep (*mastery probability*) & prediksi risiko dropout (DKT). | `Active` |
| ❓ | **Question Generator MCP** | Pembuat soal HOTS (C4–C6 Bloom), AKM Literasi/Numerasi, PISA, & Pilihan Ganda. | `Active` |
| 📚 | **Curriculum MCP** | Struktur & penyesuaian Kurikulum Merdeka, K13, Cambridge International, & IB. | `Active` |
| 🏛️ | **AI Tutor MCP (Socratic)** | Dialog Socratik pembimbing penalaran (analogi, simulasi & bimbingan bertahap). | `Active` |
| 📑 | **Research MCP** | Rangkuman paper akademis, pembuat sitasi standar APA 7th/IEEE, & literature review. | `Active` |
| 💻 | **Coding Education MCP** | Soal coding, penjelasan bug, & code review untuk 16+ bahasa pemrograman & framework. | `Active` |
| 📊 | **Learning Analytics MCP** | Generasi data Heatmap, visualisasi Knowledge Graph, & laporan analitik prediktif. | `Active` |
| 🎨 | **Content MCP** | Pembuat slide presentasi, Mindmap Mermaid.js, Flashcards, & LKPD siap cetak. | `Active` |
| 🎙️ | **Voice MCP** | Speech-to-Text, Text-to-Speech, & evaluasi pengucapan (*Pronunciation Scoring*). | `Active` |
| 🏫 | **School Admin MCP** | Otomatisasi jadwal pelajaran, sertifikat digital terverifikasi, & administrasi sekolah. | `Active` |
| 🌐 | **AI Gateway MCP** | Gateway multi-provider (Ollama Lokal, OpenRouter, OpenAI, vLLM, Claude). | `Active` |
---
## 🚀 Cara Memulai
### 1. Prasyarat Sistem
- **Python**: Versi `3.9` atau lebih baru.
- **Git**: Untuk kloning repositori.
- **Ollama** *(Opsional)*: Untuk menjalankan model AI secara *offline/lokal*.
### 2. Instalasi & Setup
```bash
# Clone repositori ini
git clone https://github.com/litera-intelligence/litera-ai-mcp.git
cd litera-ai-mcp
# Buat dan aktifkan Virtual Environment (Direkomendasikan)
python -m venv venv
venv\Scripts\activate # Windows PowerShell
# source venv/bin/activate # Linux/macOS
# Instal Dependensi
pip install pydantic requests
```
### 3. Uji Coba Server MCP via Terminal
```powershell
# Jalankan Master Educational MCP Server (57 Tools)
python litera_mcp/server.py
# Jalankan Flagship Reasoning MCP Server (4 Core Tools)
python mcp_server/litera_mcp_server.py
```
---
## 🔌 Integrasi Client MCP
Anda dapat langsung menyambungkan **Litera-AI MCP** ke aplikasi favorit Anda dengan menyalin konfigurasi di bawah ini:
### 1. Claude Desktop (`%APPDATA%\Claude\claude_desktop_config.json`)
```json
{
"mcpServers": {
"litera-ai-educational-mcp": {
"command": "python",
"args": [
"C:\\path\\to\\litera-ai-mcp\\litera_mcp\\server.py"
],
"cwd": "C:\\path\\to\\litera-ai-mcp"
}
}
}
```
### 2. Antigravity IDE / Cursor / VS Code (`mcp_config.json`)
```json
{
"mcpServers": {
"litera-ai-mcp": {
"command": "python",
"args": [
"c:/path/to/litera-ai-mcp/litera_mcp/server.py"
],
"cwd": "c:/path/to/litera-ai-mcp"
}
}
}
```
> [!TIP]
> Berkas [mcp_config.example.json](mcp_config.example.json) siap pakai sudah disediakan di repositori ini.
---
## 📁 Struktur Direktori Proyek
```text
MODEL-AI-LITERA AI/
├── 📁 litera_mcp/ # Paket Master MCP Utama (16 Modul, 57 Tools)
│ ├── 📄 server.py # MCP JSON-RPC Stdio Server Loop
│ ├── 📄 config.py # Pengaturan & Variabel Lingkungan
│ ├── 📁 domain/ # Schemas, Data Models & Pydantic Entities
│ ├── 📁 engines/ # Engine Diagnostik, DKT, & LLM Gateway
│ └── 📁 tools/ # 16 File Spesialisasi Tools MCP Pendidikan
├── 📁 mcp_server/ # Server MCP Flagship (Reasoning & Code Architect)
│ └── 📄 litera_mcp_server.py # Stdio Server Loop untuk Litera-AI Flagship
├── 📁 models/ # Penyimpanan File Model GGUF Lokal (.gguf)
├── 📄 Modelfile # Ollama Model Configuration File
├── 📄 pyrefly.toml # Konfigurasi Linter & Static Type Checker
├── 📄 mcp_config.example.json # Template Konfigurasi Client MCP
└── 📄 README.md # Dokumentasi Resmi Proyek
```
---
## 📜 Lisensi & Penulis
Proyek ini dilisensikan di bawah **[MIT License](LICENSE)** - Bebas digunakan, dikembangkan, dan didistribusikan untuk kemajuan teknologi pendidikan di seluruh dunia.
**Dikembangkan oleh:**
[](https://github.com/litera-intelligence)
<div align="center">
<sub>Built with ❤️ for Educational AI Innovation.</sub>
</div>
