LiveKit RAG Assistant
Provides async LangChain integration for building RAG (Retrieval-Augmented Generation) pipelines with semantic search and LLM-powered responses
Offers a premium glassmorphism UI for interacting with the LiveKit documentation search system, featuring real-time chat, animations, and source attribution
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
π¬ LiveKit RAG Assistant v2.0
Enterprise-grade AI semantic search + real-time web integration for LiveKit documentation
π― Features
Dual Search: Pinecone docs (3,000+ vectors) + Tavily real-time web
Standard MCP: Async LangChain with Model Context Protocol
Ultra-Fast: Groq LLM (llama-3.3-70b) sub-5s responses
Premium UI: Glassmorphism design with 60+ animations
Source Attribution: Full transparency on every answer
Related MCP server: Semantic Search MCP Server
π Quick Start
# Setup
conda create -n langmcp python=3.12
conda activate langmcp
pip install -r requirements.txt
# Configure .env
GROQ_API_KEY=your_key
TAVILY_API_KEY=your_key
PINECONE_API_KEY=your_key
PINECONE_INDEX_NAME=livekit-docs
# Terminal 1: Start MCP Server
python mcp_server_standard.py
# Terminal 2: Start UI
streamlit run app.pyApp opens at http://localhost:8501
ποΈ Architecture
Streamlit (app.py) β MCP Server β Dual Search:
ββ Pinecone: Semantic search on embeddings (384-dim)
ββ Tavily: Real-time web results
β
Groq LLM (2048 tokens, temp 0.3) β Response + Sourcesπ§ Tech Stack
Layer | Tech | Purpose |
Frontend | Streamlit | Premium glassmorphism UI |
Backend | MCP Standard | Async subprocess |
LLM | Groq API | Ultra-fast inference |
Embeddings | HuggingFace | all-MiniLM-L6-v2 (384-dim) |
Vector DB | Pinecone | Serverless similarity search |
Web Search | Tavily | Real-time internet results |
π Usage
Choose mode: π Docs or οΏ½ Web
Ask naturally: "How do I set up LiveKit?"
Get instant answer with π sources
Copy messages or re-ask from history
β‘ Performance
First query: ~15-20s (model load)
Cached queries: 2-5s
Search latency: <500ms
π οΈ Configuration
GROQ_API_KEY=gsk_***
TAVILY_API_KEY=tvly_***
PINECONE_API_KEY=***
PINECONE_INDEX_NAME=livekit-docsπ Populate Docs
python ingest_docs_quick.py # Creates 3,000+ vector chunksπ Files
app.py- Streamlit UI with premium designmcp_server_standard.py- MCP server with toolsingest_docs_quick.py- Document ingestionrequirements.txt- Dependencies.env- API keys
π¨ Troubleshooting
Issue | Solution |
No results | Try web mode or different keywords |
MCP not found | Start mcp_server_standard.py in Terminal 1 |
Slow first response | Normal (15-20s) - model initializes once |
API errors | Verify all keys in .env file |
οΏ½ Features
β Real-time chat with 60+ animations β Semantic + keyword hybrid search β Copy-to-clipboard for messages β Recent query suggestions β System status dashboard β Chat history persistence β Query validation + error handling
Version: 2.0 | Status: β Production Ready | Created: November 2025
π¨βπ» By @THENABILMAN | οΏ½ Open Source | β€οΈ For Developers
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelogβ¦
This server cannot be deployed
Maintenance
Related MCP Connectors
Ingest, manage, and retrieve documents for RAG-powered AI applications
Search your knowledge bases from any AI assistant using hybrid RAG.
Access and search Branch's official API documentation directly within any AI agent or client.
Versioned documentation registry and semantic search for AI tools and coding assistants.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to enhance their responses with relevant documentation through a semantic vector search, offering tools for managing and processing documentation efficiently.65MIT
- FlicenseNot gradedqualityDmaintenanceEnables semantic search over documentation stored in a Vectorize database using the Cloudflare Agents SDK, allowing users to query indexed content via natural language.4-
- AlicenseNot gradedqualityDmaintenanceEnables AI-powered search and querying of the Vercel AI SDK documentation using natural language questions, with tools for direct similarity search and agent-based answers.50MIT
- FlicenseNot gradedqualityFmaintenanceEnables semantic search of Weaviate documentation using vector search, providing relevant documents and code snippets.-