----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Dashboard]() * [Connecting Your Custom LLM to Vapi: A Comprehensive Guide]() * [Frequently Asked Questions]() * [Prompting Guide]() -----------------------------

### Answer It sounds like you're encountering an issue where your assistant stops responding when asked to provide location-based information, such as recommending nearby medical clinics. This might be related to how your assistant is configured to handle specific tasks and its access to external data sources like Google Maps. To address this issue in Vapi, you can integrate external APIs (like Google Maps) to fetch real-time data and enhance the assistant's functionality. However, directly implementing this within Vapi might require custom development outside of Vapi’s standard capabilities. Here's a general approach to integrating Google Maps or similar services: 1. **Use Function Calls**: Set up function calling in your assistant's configuration to invoke external services. You need to handle API calls to services like Google Maps on your server and then pass the relevant data (like clinic names and addresses) back to the assistant. Vapi allows customization and integration through server URLs where these API interactions can be managed [(1)]() . 2. **Custom LLM Integration**: If you're using a custom language model, you can customize it to interpret location-based queries and fetch data accordingly. This involves setting up a custom server to handle requests and responses between Vapi and your chosen APIs [(2)]() . 3. **Handling Data in Real-Time**: Ensure that your backend is capable of processing real-time data and interacting seamlessly with Vapi. This may involve real-time data processing and communication protocols that maintain low latency, crucial for seamless user interactions [(3)]() . For further guidance on integrating external APIs with Vapi or handling specific s

cenarios like fetching location-based data, you might consider checking out Vapi’s detailed documentation on integrating custom LLMs and function calling [(2)]() .

it will fix your issue