Introducing Firebase Studio
Firebase powers over 70 billion instances of apps daily across a wide range of platforms, from embedded systems and agentic experiences to mobile devices and web browsers, and is used by millions of developers to engage their consumers. However, full-stack development is rapidly changing, and the emergence of generative AI has changed not just the kinds of apps that may be created but also how they are constructed.
As a result, complexity increases and developers are under tremendous pressure to stay up to date with several new technologies that need human assembly. In the meanwhile, companies of all sizes are looking for methods to improve the efficiency of Artificial Intelligence app development cycles, produce high-quality software, and launch more quickly.
Google Cloud is launching a set of new features at Google Cloud Next that turn Firebase into an end-to-end platform that speeds up the whole application lifecycle. The new Firebase Studio is a cloud-based, agentic development environment powered by Gemini that has all the tools developers need to swiftly construct and publish production-quality AI apps in one location. It is now accessible to everyone in preview. Developers are able to unleash their cutting-edge, data-driven apps on Google Cloud with a number of further Firebase platform improvements. These announcements will enable developers to create new avenues for creating multi-platform AI apps.
To help creating AI apps quicker and simpler, it has introduced a number of new services in the last year, such as Genkit, Firebase’s Gemini, and Project IDX, a Code OSS fork. With the release of Firebase Studio, it is making a big leap forward by combining all of these features with Firebase services and Gemini’s creative potential to create a brand-new, natively agentic experience.
Choose from more than 60 pre-made designs for new apps, or use the App Prototyping agent to get started. Using natural language, photos, sketching tools, and screenshots, it helps you design your app’s user interface, API schema, and AI processes. Continue to refine your prototype, and when it’s ready, launch it straight to Firebase App Hosting.
Take a quick look at use and behaviour, or dive deeper into the Firebase Console for more thorough monitoring. With just one click and no further setup, you can launch your application at any moment within a Firebase Studio coding workspace. To be ready for production deployment, you can extend functionality and improve the architecture there.
Additionally, coding workspaces allow you to:
Make coding processes simpler: Gemini in Firebase will help you at every stage as you write code and test features. Perform a wide range of activities with ease, including debugging, testing, refactoring, explaining, and documenting code.
Make current apps better: Import pre-existing codebases from git-based repositories like Bitbucket, GitLab, and GitHub, or from your own computer. Make unique templates for the tech stacks you want to use and distribute them to your team.
Construct full-stack encounters: Customise and improve every part of your apps, including the database, business logic, user experience, AI model inference, agents, and retrieval augmented generation (RAG). Your AI programs may effortlessly expose and include tools like microservices and APIs.
Utilise well-known tools: To customize your workspaces, bring along your unique setups, including system tools, extensions, and environment variables. The Open VSX Registry offers access to thousands of extensions.
Adaptable choices for deployment: With built-in connectors to Google Cloud operate and Firebase backend services, you can configure your app to operate in the cloud. Deployment on your unique infrastructure is another option.
Three workspaces of Firebase Studio are now free to use during preview. Up to 30 workspaces are available to Google Developer Program participants.
Use AI agents in every step of your process
Additionally, Google Cloud giving Firebase Studio users early access to Gemini Code Assist agents. For instance, you could use Firebase Studio’s Migration agent to assist you in moving your code between Java versions, the AI Testing agent to conduct adversarial tests against AI models to find and correct potentially dangerous outputs, and the Code Documentation agent to communicate with a wiki-style knowledge base about your code to make it easier for new team members to get started.
For both automatic and manual testing, Firebase App Distribution offers a single mobile app testing solution. Real-world user interactions with your app may be replicated with Firebase App Distribution’s new App Testing agent. You may, for instance, create a test with the objective, “Find a trip to Greece.”
Using Gemini, the App Testing Agent will create a strategy to accomplish that goal and execute it on both virtual and physical devices. It will navigate your user interface and generate comprehensive pass/fail results with clear explanations and illustrations of the routes the agent took. You may test the App Testing agent on your Android app right now, and it will be available on more platforms later this year.
Develop fresh AI app experiences
Lack of standards and best practices makes integrating cutting-edge AI technologies into apps problematic. Google Cloud is investing in robust frameworks, SDKs, and tooling to help you create compelling and imaginative user experiences.
Enhanced support for more languages in Genkit
Genkit makes it easier to develop, test, and keep an eye on the AI elements of your apps. With support for multi-model orchestration, retrieval augmented generation (RAG), tool calling, human-in-the-loop interactions, Model Context Protocol (MCP), and structured output, you can create compelling agentic experiences. It is making it simpler to accomplish that in your choice language by expanding support for Go and launching early support for Python. Use Vertex Model Garden to access Gemini models, Imagen 3, and other models like Llama and Mistral. You can also use Ollama to host your own models and use community plugins to access an expanding ecosystem of third-party models.
Vertex AI creates new models in Firebase
Firebase’s Vertex AI offers a simplified, safe SDK that enables developers to include generative AI into their apps. Today, dozens of applications use it, including the offline hiking map provider HiiKER, the meal planning and shopping list management tool Meal Planner, the social media service Waveful for artists, and Life, an AI-powered diary assistant. In addition to the Gemini family of models, it also added support for Imagen 3 models (Imagen 3 and Imagen 3 Fast) in March.
This allows you to include picture production straight into your Web, Flutter, iOS, and Android applications. In order to enable additional conversational interactions in apps, such as letting users ask audio queries and receive answers, it is now providing support for the Live API for Gemini models.
Speed up contemporary, data-driven applications
With Firebase Data Connect and Firebase App Hosting, which are now widely accessible, Google Cloud also giving you more control over the architecture of your apps and the deployment procedures.
Create complex applications with Firebase Data Connect
With type-safe SDKs and quick GraphQL APIs, Firebase Data Connect provides the strong dependability of Google Cloud SQL for PostgreSQL. Create a variety of experiences, such as e-commerce systems with extensive product catalogues, social networking apps with intricate user connections, or tailored suggestions with integrated vector search.
You can now use Data Connect to:
Create schemas and queries with ease
Backend development may be greatly accelerated by using Firebase’s Gemini to automatically construct your Data Connect schemas, queries, mutations, and client SDKs.
Make use of the enhanced query capabilities
In order to assist guarantee data integrity across intricate processes, Data Connect has now increased query capabilities with native aggregation support for deeper data insights, atomic data updates, and transactions with server value expressions.
Use web frameworks while building
With created type-safe hooks and components for web frameworks, you can take advantage of tight integration and simplified data handling, facilitating the quick construction of dynamic, data-driven applications.
Use Firebase App Hosting for deployment
For contemporary, full-stack web applications, Firebase App Hosting is a git-centric, opinionated hosting option. Through the management of your app’s complete stack, including the development, CDN, and server-side rendering, App Hosting speeds up time-to-market. App Hosting takes care of the rest after you submit to GitHub. Cloud Build, Cloud Run, Cloud CDN, and other enterprise-grade Google Cloud services provide the foundation of App Hosting.
This App Hosting release allows you to:
Test and debug builds with ease
To assist you in anticipating and resolving build difficulties, App Hosting has enhanced error messages and included a local emulator.
Recover quickly from production-related events
To learn more about the health and performance of your app, use App Hosting’s new monitoring dashboard. If you notice a regression, you can quickly revert to an earlier version.
Establish a Virtual Private Cloud (VPC) connection
Allow your app to use backend services in your Google Cloud project that aren’t reachable from a public IP address (such using Cloud Memorystore to cache content or non-Firebase databases to retrieve data).
