Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp_server What vector store does the RAG pipeline use?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
An AI Engineer reference project: a Retrieval-Augmented Generation app with a governance layer, MCP exposure, and CI/CD deployment to Azure App Service.
Layer | Choice |
LLM | Groq API ( |
Orchestration | LangChain |
Embeddings |
|
Vector store | FAISS, local file-based index (no SQL/DB server) |
Frontend | Streamlit |
Governance | Presidio (PII redaction) + moderation + JSONL audit log |
Agent interop | MCP server exposing |
CI/CD | GitHub Actions → Azure App Service (no Kubernetes) |
Related MCP server: consulting-mcp-server
┌─────────────────────┐
│ Streamlit UI │
│ (app.py) │
└──────────┬───────────┘
│
┌───────────────▼────────────────┐
│ Governance Layer │
│ (PII redaction, moderation, │
│ audit logging) │
└───────────────┬────────────────┘
│
┌──────────────────────┼───────────────────────┐
│ │ │
┌──────────▼─────────┐ ┌─────────▼─────────┐ ┌──────────▼─────────┐
│ Chunking │ │ Embeddings │ │ Vector Store (FAISS) │
│ (rag/chunking.py) │ │ (rag/embeddings.py)│ │ (rag/vector_store.py)│
└──────────┬─────────┘ └─────────┬─────────┘ └──────────┬─────────┘
│ │ │
└──────────────────────┴────────────┬───────────┘
│
┌──────────▼─────────┐
│ Retriever │
└──────────┬─────────┘
│
┌──────────▼─────────┐
│ Groq LLM Generator │
└─────────────────────┘
MCP Server (mcp/mcp_server.py) exposes the same pipeline as a
`rag_query` tool for external agents/clients.python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env # then add your GROQ_API_KEYstreamlit run app.pypython -m mcp.mcp_serverpytest tests/ -vCreate an Azure App Service (Linux, Python 3.11 runtime).
Set the Startup Command to bash startup.sh.
In App Service → Configuration → Application settings, add GROQ_API_KEY
and any other values from .env.example.
In your GitHub repo, add:
Secret AZURE_CREDENTIALS (a service-principal JSON from
az ad sp create-for-rbac --sdk-auth).
Variable AZURE_WEBAPP_NAME with your App Service name.
Push to main — .github/workflows/azure-deploy.yml builds, tests, and deploys.
No SQL database — the vector index is a local FAISS file persisted to
data/vector_store/. Swap in a managed vector DB later if you need
multi-instance scaling.
No Kubernetes — deployment target is Azure App Service (PaaS), which is simpler to operate for a single-container Streamlit app. Move to AKS only if you need pod-level autoscaling or a multi-service mesh.
This server cannot be deployed
Ingest, manage, and retrieve documents for RAG-powered AI applications
Cloud or self-hosted knowledge for AI agents: hybrid search, reranking, GraphRAG, scoped MCP tools.
Medical RAG: semantic search for clinical guidelines, drug interactions, diagnoses & EHR data.
Medical RAG: semantic search for clinical guidelines, drug interactions, diagnoses & EHR data.
