3GPP MCP Server
Enables AI assistants to access and search 3GPP telecommunications specifications through direct integration with the TSpec-LLM dataset. Provides real-time specification content, implementation requirements, and multi-spec comparisons for 3GPP standards development.
README
3GPP MCP Server V3.0.0 - Direct Specification Access
Transform your AI assistant into a 3GPP specification expert with direct access to TSpec-LLM's 535M word dataset!
What This Does
Before: Ask AI about 3GPP specifications - Get generic responses based on training data After: Ask AI + 3GPP MCP Server V3.0.0 - Get direct access to current specification content with structured, agent-ready responses
Revolutionary V3.0.0 Architecture
V3.0.0 represents the True MCP approach - lightweight API bridge providing direct specification data:
Agent Query → MCP Tools → External APIs → Real Specification Data
Key Benefits:
- True MCP Architecture - Lightweight API bridge (~10MB vs 15GB+)
- Sub-500ms responses - Intelligent caching with external API integration
- Agent-optimized - Structured JSON responses for AI agent consumption
- Real specification data - Direct access to TSpec-LLM's 535M word dataset
- External API integration - Hugging Face + 3GPP.org APIs
- Infinite scalability - Stateless API calls, no local storage limits
Quick Start (30 Seconds!)
Direct MCP Setup (Recommended)
Claude Desktop users:
claude mcp add 3gpp-server npx 3gpp-mcp-charging@latest serve
For other MCP clients: Add this to your MCP configuration:
{
"mcpServers": {
"3gpp-server": {
"command": "npx",
"args": ["3gpp-mcp-charging@latest", "serve"],
"description": "3GPP MCP Server - Direct access to TSpec-LLM and 3GPP specifications",
"env": {
"HUGGINGFACE_TOKEN": "optional-for-enhanced-access"
}
}
}
}
Alternative: Auto-Configuration
# One-command installation with auto-configuration
npx 3gpp-mcp-charging@latest init
# Client-specific installation
npx 3gpp-mcp-charging@latest init --client claude
npx 3gpp-mcp-charging@latest init --client vscode
npx 3gpp-mcp-charging@latest init --client cursor
Test It Works
Ask your AI assistant: "Search for 5G CHF implementation requirements in TS 32.290"
You should get structured specification content with implementation guidance, dependencies, and testing considerations!
Available Tools (V3.0.0)
| Tool | Purpose | Input | Output |
|---|---|---|---|
search_specifications |
Direct TSpec-LLM search | Query + filters | Structured spec results + relevance scores |
get_specification_details |
Comprehensive spec details | Specification ID | Full metadata + implementation guidance |
compare_specifications |
Multi-spec comparison | Array of spec IDs | Comparison matrix + migration analysis |
find_implementation_requirements |
Requirements extraction | Spec scope + focus | Technical requirements + testing guidance |
Example Queries
Direct Specification Search:
"Find charging procedures in 5G service-based architecture"
→ Returns: TS 32.290 excerpts, CHF implementation details, Nchf interface specifications
Implementation Requirements:
"Extract implementation requirements for converged charging in Release 17"
→ Returns: Technical requirements, dependencies, testing considerations, compliance notes
Specification Comparison:
"Compare charging evolution from TS 32.240 to TS 32.290"
→ Returns: Evolution timeline, migration analysis, implementation impact assessment
What You Get
Direct Specification Content
- Real-time access to TSpec-LLM's comprehensive 3GPP dataset
- Structured content excerpts with relevance scoring
- Official specification metadata integration
Agent-Ready Responses
- JSON-formatted responses optimized for AI agent consumption
- Consistent schema across all tool responses
- Rich metadata embedded in all responses
Implementation Intelligence
- Technical requirements extraction from specifications
- Dependency analysis and implementation guidance
- Testing considerations and compliance mapping
Performance Benefits
- <500ms cached response times
- <2s fresh API call responses
- <10MB memory footprint (stateless design)
- Unlimited concurrent users (external API scaling)
Architecture
Core Components
External API Integration Layer
- TSpec-LLM Client: Direct integration with TSpec-LLM dataset via Hugging Face APIs
- 3GPP API Client: Integration with official 3GPP.org APIs for metadata
- API Manager: Unified orchestration layer for all external APIs
MCP Tool Layer
- search_specifications.ts: Direct specification search implementation
- get_specification_details.ts: Comprehensive specification details
- compare_specifications.ts: Multi-specification comparison
- find_implementation_requirements.ts: Requirements extraction
Caching & Performance
- NodeCache: Intelligent API response caching
- Rate Limiting: Respectful external API usage
- Error Handling: Robust API integration with fallbacks
Project Structure
3gpp-mcp-server-v2/
├── src/ # V3.0.0 source code
│ ├── api/ # External API integration layer
│ │ ├── tspec-llm-client.ts # TSpec-LLM Hugging Face client
│ │ ├── tgpp-api-client.ts # 3GPP.org official API client
│ │ ├── api-manager.ts # Unified API orchestration
│ │ └── index.ts # API exports
│ ├── tools/ # MCP tool implementations
│ │ ├── search-specifications.ts # Direct specification search
│ │ ├── get-specification-details.ts # Comprehensive spec details
│ │ ├── compare-specifications.ts # Multi-spec comparison
│ │ ├── find-implementation-requirements.ts # Requirements extraction
│ │ └── index.ts # Tool exports
│ ├── types/ # TypeScript interfaces
│ └── index.ts # MCP server implementation
├── bin/ # CLI installation tools
├── docs/ # Documentation
├── tests/ # Test suite
└── package.json # NPM package configuration
Requirements
- Node.js 18+ - Download from nodejs.org
- MCP-compatible AI assistant (Claude Desktop, VS Code, Cursor, or others)
- Internet connection - For external API access
- Optional: Hugging Face token - For enhanced API access
Installation Options
Option 1: Direct MCP Configuration (Recommended)
No local installation needed! Server runs directly from NPM.
Option 2: Development Setup
# Clone and setup for development
git clone <repository-url>
cd 3gpp-mcp-server/3gpp-mcp-server-v2
npm install
npm run build
npm run start
Option 3: Auto-Configuration
npx 3gpp-mcp-charging@latest init
Environment Variables
# Optional: Enhanced API access
export HUGGINGFACE_TOKEN="your-huggingface-token"
# Optional: Custom cache settings
export CACHE_TIMEOUT="3600" # seconds
export ENABLE_CACHING="true"
Version Evolution
| Version | Approach | Storage | Architecture |
|---|---|---|---|
| V1 | Data Hosting | 15GB+ local dataset | Heavy, non-MCP compliant |
| V2 | Guidance Templates | <100MB knowledge base | Lightweight, guidance-only |
| V3.0.0 | Direct Data Access | <10MB (stateless) | True MCP API bridge |
Development
Available Scripts
npm run build # Build TypeScript
npm run dev # Development with watch
npm run start # Run the server
npm run test # Run tests
npm run lint # Lint code
npm run clean # Clean build artifacts
Adding New Tools
- Create tool class in
src/tools/ - Define tool schema with input/output types
- Implement
execute()method with API integration - Export tool and register in
src/index.ts
API Integration
- Extend
TSpecLLMClientfor new TSpec-LLM capabilities - Extend
TGPPApiClientfor additional 3GPP.org endpoints - Add orchestration methods to
APIManager
Contributing
Contributions welcome! Please focus on:
- API integration improvements
- Performance optimizations
- New MCP tool implementations
- Documentation enhancements
License
BSD-3-Clause License - see LICENSE file for details.
Acknowledgments
Research Foundation
This project's V3.0.0 architecture was fundamentally inspired by the TSpec-LLM research:
TSpec-LLM: A Large Language Model for 3GPP Specifications
- Paper: https://arxiv.org/abs/2406.01768
- Authors: Rasoul Nikbakht, et al.
- Dataset: TSpec-LLM on Hugging Face
Originally planned as a document reference MCP, discovery of the TSpec-LLM research paper fundamentally changed our approach. The paper's demonstration of significant accuracy improvements (25+ percentage points) through direct LLM access to 3GPP specifications convinced us to pivot from document hosting to external API integration with their comprehensive 535M word dataset.
Technical Foundation
- Built using the Model Context Protocol SDK
- Integrates with TSpec-LLM dataset
- Supports 3GPP specifications from 3GPP.org
V3.0.0: True MCP architecture providing direct specification access through external API integration.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.