Cymbal Air Toolbox Demo
A demonstration-only, Google-published reference implementation of a customer-service agent for a fictional airline, using LangGraph, tools and retrieval-augmented generation to query Google Cloud databases through MCP Toolbox.
This project is a production-quality reference implementation, explicitly for demonstration and not an officially supported Google product, showing how to build an agentic application that answers questions about a fictional airline (Cymbal Air) and its hub airport, such as which shops are near a gate or which flights head to a given city.
The demo combines retrieval-augmented generation with agent-based orchestration to query data stored in Google Cloud databases (AlloyDB, BigQuery, Spanner and PostgreSQL are named among its topics), reaching that data through Google's MCP Toolbox rather than direct database access, so the agent's responses stay grounded in retrieved data instead of relying purely on the model's training.