Today’s platform update is a huge leap forward for Losant's AI-powered development features by introducing application editing capabilities to our MCP server.
When we introduced our Model Context Protocol server back in June, we did so knowing that those initial read-only capabilities were just a first step towards our goal of IoT applications developed using AI and large language models. Today's update is a major milestone towards that goal: Using your LLM of choice and connecting it to Losant's MCP server, IoT solution developers can author workflows, configure experience endpoints, generate dashboards, and provision fleets of devices with a few natural language prompts.
As an example, users can create a new, blank Losant application and then provide a prompt like the following to Google Gemini, ChatGPT, Claude, or any other LLM agent to which they've connected Losant's MCP server:
In my "GPS Tracker" application, build out the resources I need for the Digital Matter Barra GPS tracker. I need a device recipe for provisioning new devices, a webhook and a workflow for processing data sent to the application by Digital Matter's platform, and a dashboard for showing a device's movement history and current status.
The model then looks up the hardware's specifications and the Digital Matter API documentation and creates the necessary resources:
We can then review the work in our application and ask for updates as needed:
Please modify the workflow to validate incoming payloads fit the expected schema, and also have it automatically register new devices from the recipe if a corresponding device is not found in the Losant application.
Finally, we can use the MCP integration to build a data simulator workflow to test the agent's work:
Now, please create a separate workflow to simulate one of these trackers. Have it report data to the webhook every 5 seconds as it is traveling from Cincinnati, OH to Mason, OH along Interstate 71.
And with a few final tweaks, we have a fully functional Losant application with dynamic device registration, third-party data ingestion, telemetry collection, and data visualization!
We can then continue developing this into an end-user-facing application, prompting the MCP server to create experience endpoints and pages and build out our multi-tenancy model based on the tags from each dynamically registered device.
We're not finished yet. Through the rest of the year we'll be integrating these generative AI capabilities directly into the Losant user interface, honoring our longstanding commitment to usability and developer experience as top priorities. We're tremendously excited about these updates and know our IoT application developers will be as well.
As always, this release comes with other features and improvements, including:
With every new release, we listen to your feedback. By combining your suggestions with our roadmap, we can continue to improve the platform while maintaining its ease of use. Let us know what you think in the Losant Forums.