California Housing MCP Agent
Enables California housing price predictions, batch scoring, and agent-based chat using a trained ML model served via MCP tools.
README
California Housing MCP Agent
Practica end-to-end de MLOps con el dataset California Housing. El proyecto entrena modelos de regresion, registra el mejor en MLflow, sirve predicciones con FastAPI, expone herramientas MCP y ofrece un frontend HTML para inferencia individual, batch y chat con agente.
Componentes
training/train.py: entrena candidatos, evalua metricas y registra el modelo ganador en MLflow.api/main.py: aplicacion FastAPI con lifespan, CORS, observabilidad basica y servidor MCP embebido opcional.api/routes/v1/: endpoints de health, prediccion, batch, metadatos y chat del agente.agent/runtime.py: runtime del agente conectado a MCP mediante OpenAI Agents SDK y LiteLLM.frontend/index.html: interfaz local para probar predicciones y batch scoring.
Requisitos
- Python 3.11 o superior.
uvinstalado.- Una clave de Gemini si
FREE_MODEL=true, o una clave de OpenAI siFREE_MODEL=false.
Configuracion
uv sync
Copy-Item .env.example .env
Edita .env y define solo las claves que vayas a usar:
GEMINI_API_KEYcuandoFREE_MODEL=true.OPENAI_API_KEYcuandoFREE_MODEL=false.
Entrenar y registrar el modelo
En una terminal, arranca MLflow:
uv run --env-file .env -- mlflow server --backend-store-uri sqlite:///mlflow.db --default-artifact-root ./mlruns --host 127.0.0.1 --port 5000
En otra terminal, ejecuta el entrenamiento:
uv run --env-file .env -- python training/train.py
El script compara candidatos, registra el modelo si mejora el historico y asigna el alias configurado en MLFLOW_MODEL_ALIAS.
Ejecutar la API
Con MLflow disponible y un modelo registrado:
uv run --env-file .env -- uvicorn api.main:app --reload --host 127.0.0.1 --port 8000
Endpoints principales:
GET /api/v1/healthPOST /api/v1/predictPOST /api/v1/batch-predictGET /api/v1/model-infoGET /api/v1/schemaPOST /api/v1/agent/chatPOST /mcpcuandoENABLE_MCP_SERVER=1
La documentacion OpenAPI queda disponible en http://127.0.0.1:8000/docs.
Frontend
Abre frontend/index.html en el navegador y usa http://127.0.0.1:8000/api/v1 como base URL. La pantalla permite:
- prediccion individual,
- subida de CSV para batch scoring,
- chat con el agente,
- inspeccion de request/response JSON y metricas de entrada.
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.