AzureFeeds

Describe the app you want using Rayfin’s open-source SDK with GitHub Copilot, and generate your full backend in code — schemas, relationships, and access policies included. Deploy to Microsoft Fabric with a single CLI command and immediately inherit enterprise data security, identity controls, and audit compliance already in place across your data estate. 

Connect your app’s live operational data to years of historical records in Fabric from the moment you deploy, no pipelines, no data movement. Query across both datasets using a Fabric data agent you spin up directly on your app’s data. 

Will Thompson, Microsoft Fabric Principal Product Manager, shares how to take an app from idea to governed production deployment in a single session.

Describe the app you want. 

Rayfin and GitHub Copilot generate your full backend in code- schemas, relationships, and access policies. No manual wiring. See how it works.

One CLI command. 

Fully managed SQL database, APIs, and static hosting in Microsoft Fabric. Rayfin handles provisioning automatically. Take a look.

Spend less time configuring. Spend more time creating.

Get managed backend services, authentication and authorization, data APIs, built-in analytics, and AI-ready architecture with Rayfin. Check it out.

QUICK LINKS: 

00:00 — Simplify backend complexity 

01:20 — Home delivery service app 

01:48 — Data analysis app 

02:26 — See the build experience 

03:08 — Copilot Generates Full Backend 

03:47 — Authorization defined alongside schema 

05:06 — One CLI Command Deploys to Fabric 

05:21 — Create analytics app & add pages 

06:31 — App Data Connects to Fabric Data Estate 

06:55 — Conversational Data Agent on App Data 

08:13 — Wrap up

Link References 

Get started at https://aka.ms/rayfin

Unfamiliar with Microsoft Mechanics? 

As Microsoft’s official video series for IT, you can watch and share valuable content and demos of current and upcoming tech from the people who build it at Microsoft. 

• Subscribe to our YouTube: https://www.youtube.com/c/MicrosoftMechanicsSeries 


• Talk with other IT Pros, join us on the Microsoft Tech Community: https://techcommunity.microsoft.com/t5/microsoft-mechanics-blog/bg-p/MicrosoftMechanicsBlog


• Watch or listen from anywhere, subscribe to our podcast: https://microsoftmechanics.libsyn.com/podcast

Keep getting this insider knowledge, join us on social: 

• Follow us on Twitter: https://twitter.com/MSFTMechanics


• Share knowledge on LinkedIn: https://www.linkedin.com/company/microsoft-mechanics/


• Enjoy us on Instagram: https://www.instagram.com/msftmechanics/


• Loosen up with us on TikTok: https://www.tiktok.com/@msftmechanics 

Video Transcript:

-What if building a real, production-ready enterprise app didn’t take months, but just a few hours? Today, AI can generate app logic and front ends almost instantly. And while everything works great with test data, the moment you try to connect that app to real enterprise data, everything slows down. Because now you have to wire up a database with access control, set up authentication, configure backend services, and make sure everything is secure and scalable. Your quick app turns into a complex backend project with multiple failure points, and that’s where Rayfin comes in. It’s designed to simplify backend definition in code and elevate the developer experience. Our open-source SDK and CLI grounds your agentic coding tools so that when you describe the app that you want to build, it can automatically define your app’s backend in code, including your data schemas, relationships, and access policies. 

-And then with a single CLI command, you deploy to our unified data and analytics platform, Microsoft Fabric. This spins up a fully managed backend with a SQL database with familiar billing and sharing models, as well as static hosting. APIs are generated automatically to ensure your app can communicate securely to data and services and Fabric’s identity and data security layer ensures that when your app accesses or queries data, it respects existing policies in place across your data estate. 

-To show how this works in action, I’ll start by showing two apps I built with Rayfin and then walk you through how I created them. My first app is a mobile home delivery service app for a home improvement retailer, with all orders delivered directly to customers’ doors. Drivers must be signed in to access the app, ensuring that every delivery is securely tied to an authenticated user. They can record deliveries, capture photos as proof, and even collect customer signatures to acknowledge receipt. And all this happens in real time. 

-Next, as a business, we want to learn from the data captured in our delivery app. So I’ve built a second app for data analysis. Because it runs in Fabric, the operational data is already connected to our broader enterprise data estate. It has access to years of customer data in Fabric, modeled and ready to use. And by connecting that data to the live data that our delivery app captures, retail managers now have immediate insight into how the new delivery service is impacting customer satisfaction. Now these apps were built in minutes. They’re not a proof of concept. They have real auth, real data, and real governance, and they work directly with enterprise data already in Fabric. 

-Let me walk you through the build experience. I’ll open a terminal and run the Rayfin create command, and this will walk me through my entire project setup. I can see that I have various templates that I can use to get started. So let’s go with the blank app template. It’s asking for a name and I’ll name it Delivery App. And you can see it’s creating the project. And under the covers, it’s installing all the necessary dependencies and adding the files I need, including the agent instructions for using the SDK and CLI. Now I can CD into my project and start the development server. Since deploying to Fabric, it prompts for a Fabric workspace name and I’m going to use Zava. And the deployment process is now in full swing with Fabric provisioning my foundational resources. 

-So now with our Rayfin foundation in place, let’s build the app. Using VS code with GitHub Copilot, I can now describe the app I want and Copilot knows how to build it, including the backend. Let’s try it out, I’ll type a prompt, build a Rayfin delivery app where drivers log deliveries with photos and item conditions and customers confirm receipt. I’m also going to start the Rayfin MCP server and this will help ground Copilot in all the latest best practices for Rayfin. As Copilot is getting to work, we can see it’s using the documentation for the MCP and CLI tool to understand the specifics of Rayfin. And Copilot will run independently using the skills defined by the Rayfin template. That’s going to take a few minutes, so we’re going to jump ahead to the results. 

-Here’s the app right here in code and there’s lots more to it, in fact, to understand the depth of what was created, let’s take a more detailed look. Inside the Rayfin folder, we can see our whole data schema, customers, deliveries, items, photos, drivers, and more. And if we take a look at how Copilot defined the driver entity, this is where Rayfin shines. The schema lives right in the code, which means instead of keeping the rules about your data in separate places like databases, SQL files, or configs, they’re built right into the code. So things are less likely to break and easier to keep consistent. Each field maps to a column in the database. ID name, email, and optional phone number with types and constraints built in. 

-And it captures relationships too. It calls out a driver has many deliveries. But here’s the key part, the access policy sits right next to the data. This rule says a user can only ever touch their own driver record. Authorization is defined alongside the schema not bolted on later. And if we open the delivery entity, you can see some of the scale Rayfin is capable of with richer relationships, defined with drivers, customers, and items, and a smarter policy where delivery can be read by an authenticated driver or its customer. And of course I can always customize any of this by hand if I need to. So that was the build experience. 

-Now let’s look at how we can deploy the app to Microsoft Fabric. I’ll ask Copilot to deploy our app and Copilot is using the Rayfin Up command to deploy everything we’ve defined straight to Fabric. And just like that, our app is deployed and available. Next, directly from Fabric, I’ll show you the steps to create our analytics app. I’ll start in the Fabric portal this time and I’ll call this app Delivery Analysis. This leverages a template that the Microsoft Power BI team built. It’s called a Data App template, and it gives our coding agent what it needs to connect to semantic models and create beautiful interactive visuals. 

-So I’ll select Data App, and now I can manually follow these steps or just copy the prompt and run it in VS Code to get started scaffolding my project. I’ll move over to VS Code and paste in the prompt and the agent knows exactly how to start up my new Rayfin app. Once it’s finished, you’ll see that the delivery analysis directory has been added. Now we can also add pages to our new app using Copilot. Here I want to add an insights page. So I’ll prompt, Add a customer satisfaction page, which shows historical CSAT scores from our linked semantic model and compare it against a given driver’s delivery ratings. The important detail here, I’m pointing it straight at our Fabric semantic model, so it’s using a preexisting and trusted business model. This process will run for several minutes, creating and executing a plan and querying our semantic model for what it needs. 

-To save time, I’ll jump back in near its completion. Now the app is ready and I can view it from the provided link. And we can see historical customer satisfaction data trending up since we launched the delivery service with details broken down by driver. Now, data isn’t moved between systems. Everything stays connected from the start. The backend is already connected to your enterprise data estate. And because all of your new data is now also landing in Fabric, there’s lots you can do on top of it. 

-Now that our app is live and generating data, the next step is making that data actionable through conversation. So instead of building another report or connecting the data to a Fabric notebook, I’m going to go ahead and create a data agent directly on top of this app’s data. I’ll name the agent, Zava Insights Agent, and next I’ll point it at the warehouse I configured earlier. This combines my operational database and historical data as the source. And now I can ask questions against my data just like I would to an analyst. 

-Let’s start with a simple question. I’ll ask, what is the average customer rating across all our deliveries? The agent understands the data model and returns the answer instantly as 4.56 out of five. But now I’ll go deeper. How does our customer satisfaction compare before and after we launch the new delivery service? It doesn’t just retrieve the data. It reasons across time and trends in the model. And we can see that it’s gone from 3.8 to 4.4 out of five since launching the new service. So instead of navigating dashboards, the business can now ask questions directly against the app and get answers grounded in the same governed data that powers the app. 

-And just like any Fabric artifact, I can share the app that we created earlier with other people in my organization. So let’s grant Lynne access. They were asking about the delivery analysis and now they can explore the data themselves without waiting on reports. Rayfin helps you go from idea to production faster. By using AI to build and deploying to Fabric, your app inherits existing data security controls, permissions and audit requirements, keeping everything compliant and under control. 

-To learn more, check out aka.ms/rayfin and keep watching Microsoft Mechanics for the latest tech updates and thanks for watching.