Inspiration
Travel plans can quickly fall apart when the weather changes, local conditions shift, or finding the right places becomes difficult. At the same time, voice assistants in India do not always provide the natural Hindi and Marathi experiences that local users need.
We wanted to build a smarter, more adaptive city guide that helps people explore their cities, discover places, and make better travel decisions through a conversational experience. This idea led us to CityMind Voice Concierge, designed with Maharashtra, India, as our first launch region.
What it does
CityMind Voice Concierge is a self-hosted Model Context Protocol (MCP) server designed to connect city-exploration capabilities with Alexa+ experiences.
It exposes five tools that allow an AI assistant to interact with city information and travel-related functionality:
City Discovery: List supported cities and help users begin exploring.
Place Search: Search for places relevant to a user's interests and needs.
Condition Awareness: Check relevant real-world conditions to support more informed travel decisions.
Conversational Assistance: Enable an assistant to use available tools to respond to city-exploration requests.
Extensible Integration: Provide a foundation for expanding city-guidance capabilities through the MCP interface.
Our goal is to make exploring a city more convenient, context-aware, and accessible, particularly for users who prefer interacting through voice.
How we built it
We built CityMind Voice Concierge around the Model Context Protocol, using a self-hosted MCP server with Streamable HTTP transport. This architecture exposes city-exploration functionality as tools that a compatible AI assistant can discover and invoke.
Rather than treating the assistant as a simple question-and-answer interface, we designed the system to let it call tools for relevant information when responding to a user's request.
The project focuses on bringing together conversational AI, city information, and real-world conditions in a foundation that can support a more useful travel experience. Maharashtra is our initial target region, with Hindi and Marathi voice interactions forming an important part of our vision.
Challenges we ran into
One of the central challenges was designing a useful city-guide experience that goes beyond generic recommendations. Travel decisions depend on changing conditions, so the system needs to support access to relevant information rather than relying entirely on static responses.
Another challenge was structuring the functionality as MCP tools so that an AI assistant can interact with it through a clear interface. We also considered the needs of users in India, where language accessibility and locally relevant information are important to making voice assistants more useful.
Accomplishments that we're proud of
We are proud to have built CityMind Voice Concierge, a self-hosted MCP server designed to bring city exploration capabilities into an Alexa+ experience.
One of our key accomplishments was structuring city-related functionality into five tools, making it possible for a compatible AI assistant to interact with the system through a defined interface. We focused on combining conversational AI with place discovery and real-world condition awareness to make travel planning more useful.
We are also proud of designing the project with Maharashtra as our initial target region and keeping Hindi and Marathi accessibility in mind. This project represents an important step toward building AI-powered city experiences that are locally relevant, adaptable, and accessible through natural conversation.
What we learned
Building CityMind Voice Concierge helped us understand how the Model Context Protocol (MCP) can connect AI assistants with external tools and services.
We learned how a self-hosted MCP server using Streamable HTTP can expose structured capabilities to compatible AI clients. We also learned the importance of designing clear tool interfaces, separating information retrieval from conversational responses, and considering changing real-world conditions when building travel-related applications.
Beyond the technical implementation, this project strengthened our understanding of user-focused AI design. We realized that a useful city assistant should do more than provide answers: it should help people discover places, understand relevant conditions, and make better-informed decisions.
What's next for CityMind Voice Concierge
Our next goal is to make CityMind Voice Concierge a more comprehensive and reliable companion for exploring cities.
We plan to expand city coverage beyond our initial Maharashtra focus, improve access to timely local information, and enhance the system's ability to support travel decisions when conditions change.
We also want to develop more natural Hindi and Marathi conversational experiences, improve the relevance of place recommendations, and explore additional integrations that make city exploration easier through voice.
Our long-term vision is to create an AI-powered city guide that combines local knowledge, real-world context, and conversational assistance to help people explore their surroundings with greater confidence and convenience.
Built With
- ai
- alexa+
- artificial-intelligence
- mcp
- python
- streamable-http
- voice-assistant
