About the Project

The idea for FieldCall came from attending an outdoor event and thinking about how much the weather can affect an experience people have planned, travelled for, and looked forward to. A general weather forecast might tell you what the weather will be like, but it does not always answer the question people actually care about: will the conditions be suitable for the event I am attending?

I realised this was not just a problem for one event or one person. Many people make plans for outdoor gatherings, concerts, celebrations, and community events without an easy way to compare the forecast with the weather conditions that matter to them. I wanted to build something that could help people make more informed decisions before setting out.

That is what inspired me to build FieldCall. It allows users to enter an event address or coordinates, choose their own temperature, wind, and precipitation limits, and compare those limits against a source linked forecast for the event's location and time window. Each check is saved with its source and timestamp, so users can review what the forecast showed. FieldCall provides information to support planning, not a guarantee of safety.

I built FieldCall around an MCP server using Streamable HTTP, connecting the web experience to tools for saving events, retrieving forecasts, evaluating weather conditions, and reading stored results. I chose a deterministic approach because comparing forecast data against user defined limits does not need unnecessary AI. It needs clear rules, traceable data, and honest results.

One of the biggest challenges was accounting for unavailable integrations and differences in forecast coverage without pretending that missing information was a successful result. I also wanted the tool to support locations beyond the United States. Users can enter coordinates manually when Google address lookup is not configured, although live integrations still need end to end verification with the required credentials and services.

Building FieldCall taught me that a useful product is not just about showing information. It is about helping people understand where that information came from, when it was retrieved, and what it can actually tell them. I hope FieldCall can help more people think ahead about the weather before an outdoor event, rather than discovering too late that the conditions are different from what they expected.

Built With

  • alexa+-simulated-web-experience
  • css
  • docker
  • google-maps-geocoding-(optional)
  • google-oauth
  • html
  • javascript
  • met-norway-locationforecast-api
  • model-context-protocol-(mcp)-over-streamable-http
  • python
  • sqlite
Share this project:

Updates

Submission history