Chattende Fahrplan

About

Inspired by the 1996 "Sprechender Fahrplan" by SBB, this experimental Spring AI / MCP-compatible server provides real-time Swiss railway timetable information to language models. It enables LLMs to query the SBB Journey API directly using a set of specialized tools for planning journeys, listing connections, and retrieving raw data. Designed for 'bring-your-own LLM' scenarios, it demonstrates how LLMs can perform real-time data queries, combine various result formats, and reason over structured SBB journey data within the Model Context Protocol framework. It has been tested with various LLMs, including OpenAI OSS GPT, Apertus 8B, Qwen-3 8B, Mistral Large, Claude Sonnet, and GPT-5.

Key Features

• Experimental study project demonstrating LLM interaction with structured data
• Dedicated LLM tools for next journey, specific journey planning, and connection listings
• Real-time SBB Journey API integration for Swiss railway timetable queries
• Compatibility with various LLM runtimes and frontends (LM Studio, Le Chat, Claude, ChatGPT)
• Deployable locally (Java 21) or on Google Cloud Platform (GCP Cloud Run)
• 5 GitHub stars

Use Cases

• Integrating real-time SBB transportation data into custom LLM applications or assistants
• Querying Swiss railway timetables and planning journeys using natural language prompts with LLMs
• Exploring and demonstrating LLM capabilities for reasoning over structured data within a transportation context