The demand for AI-powered features in applications is exploding, but integrating generative AI is often complex and slow. This session introduces Firebase Studio, a new AI powered Cloud Based Workspace on the Firebase platform designed to dramatically simplify building AI backends.
Join us for a demo-driven guide on leveraging the Gemini API to build sophisticated features with remarkable speed. You’ll leave with a clear understanding of how to use the Firebase ecosystem to rapidly prototype and ship scalable, secure, and intelligent applications
Sobre Rody Davis: Developer Advocate for Firebase Studio at Google. Working closely with Flutter, Angular, Firebase, Go and Gemini in Google Cloud. In my time at Google I have also supported Material Design, Lit (previously Polymer) and Google Developers (Solutions).
I have a background in systems administration, audio engineering, music and electrical engineering. I am very passionate about open source and creative applications of software. Currently I am working with teams to take advantage of AI in software development and integration into apps.
My mission is to make the intersection of Google products work better together.
🛤️ Track: DEVELOPMENT 📋 Formato: Charla
0:00 · Working with AI can vary a lot depending on your skill level, experience, and model provider, or even the AI agents or tools you’re using. At Google, we have many AI offerings ready to meet you where you are and help you take your next or current project to the next level. Here at Firebase, we believe that to make great apps, even through the assistance of AI, will require making it easy to build with agents that can provision resources, author backing code, and build complex UIs.
0:31 · So, who here has heard of Firebase?
0:33 · Raise your hand. Wow, that’s awesome.
0:36 · What about Firebase Studio?
0:39 · Nice. Uh, Firebase AI logic or Genkit.
0:43 · Okay, less people, but still cool.
0:45 · Firebase is Google’s platform to build, prototype, deploy, and run modern AI powered experiences that users love.
0:53 · What makes fair Firebase so special?
0:55 · Well, you know it very well. First, it’s a comprehensive platform. And second, it’s very easy to get started. We will give you a well-lit path to accomplish your objectives. There’s no need to hack together a solution from multiple blog posts and forums or different parts of the internet. But we don’t stop there.
1:14 · Firebase is backed by GCP’s infrastructure, security, and reliability, and trusted by millions of businesses. And most importantly, it’s extensible and flexible. You get to decide which parts of the platform make the most sense for your application and you bring in the other tools and services as you need.
1:35 · We recently announced Firebase AI logic in the as the next evolution of Vertex and Firebase. We recognize that integrating AI into your applications can be challenging. A year ago, we launched Vertex and Firebase making a lot easier and is already used by thousands of customers in production.
1:54 · For example, who’s accidentally shipped their API key to the client? I know there’s at least one.
2:02 · Firebase AI logic is one of the easiest ways to integrate, innovate, and deploy AI seamlessly into your apps.
2:09 · On your client, you can use Firebase AI logic proxy to securely access Google AI and vertex models such as Gemini and Imagine. And on the server, you can use Genkit. Firebase AI logic is integrated with other Firebase services like app check remote config and AI observability and remote config.
2:32 · In addition to the Vertex API, you can use the Firebase AI logic client SDKs to connect to the Gemini developer API which has a free tier. And last of not not least, we’re launching Unity support which also includes support for Android XR.
2:49 · We recently updated Firebase AI logic to have an experimental mode for using hybrid inference. This is currently offered on Chrome on desktop. With hybrid inference, your inference will happen on the device if the ondevice is available. Otherwise, it will seamlessly fall back to cloud hosted models.
3:09 · We are also expanding AI monitoring.
3:12 · Previously only available to Genkit, you could now monitor your AI logic SDKs coming from the client in a single pane of glass where you can see things such as latency, error rates, token use, or even stack traces. And it’s very useful.
3:26 · But we didn’t stop there.
3:28 · Genkit is also now a part of Firebase AI logic. Genkit, for those that don’t know, is our open- source toolkit for building powerful AI backends, offering libraries, tools, and out-of-the-box integrations with many popular models and cloud services.
3:44 · You’ve given us great feedback on Genkit, and I’m happy to say that we’ve heard you loud and clear. At Nex, we officially launched support for Python in Genkit, and since then, we’ve also been hard at work on supporting and improving Go. But we’ve also introduced dynamic model registration, which means you no longer need Genkit to have an update to leverage the latest models, and it can just seamlessly update and have access.
4:08 · We’ve recently announced genkit.dev, which you definitely need to check out because it’s our go-to single surface for the most up-to-date documentation and code examples on how to use Genkit effectively in your applications.
4:24 · Now, almost every developer needs a database and you’ve been asking for Firebase to have a relational database forever when we’ve heard you loud and clear. And just over a year ago at IO, we launched Firebase data connect.
4:38 · Firebase data connects to your application to your data in Postgress Cloud SQL through secure serverless APIs.
4:46 · We built data connect with one premise.
4:48 · You write the query and we do the rest.
4:51 · That means that all you have to do when you write the query is we’ll create an API endpoint, define an O policy for that endpoint, generate uh strongly typed client SDKs in your favorite languages like Noode.js, Dart, Cotlin, and Swift, and it will also create tool definitions so you can pass those queries directly to your LLM that you’re using. This significantly accelerates app development and reduces time to value.
5:18 · We also recently supported uh releasing Firebase data connect to go to GA at cloud next and we’ve also added two things that people have been really excited about. One is advanced query improvements so you can see really nice insights into the sophisticated data retrieval in transactions from additional server values and multi-step capabilities.
5:38 · We’ve also introduced a schema and query generation agent which is really amazing for when you’re wanting to build your application and you give it a list of the models that you’re trying to build and we’ll generate the schema for you. And this is also available inside Gemini code assist and VS code extensions.
5:57 · Next, let’s talk about Firebase app hosting. Firebase app hosting is an opinionated serverless hosting solution for modern full stack frameworks like Nex.js. These fullstack frameworks make it have simpler to get the benefits of serverside rendering so you can have improved SEO and faster load times with the ease and familiarity of JavaScript.
6:17 · But this means that deploying the web isn’t as simple as throwing up a bunch of static files on a CDN. You also need a back-end server. You need serverside rendering a build pipeline for the static generation. And of course, you also need a CDN for the caching. There’s not a single part of this that you can do without, but with Firebase hosting, we’re here to make the entire stack simpler for us.
6:38 · App hosting seamlessly integrates with the deployments from GitHub to Google Cloud, so it sets up all the necessary cloud services and detects what you need in your source code. App hosting automatically builds the assets in Cloud Build, deploys the dynamic content to Cloud Run, and serves the cached assets from Cloud CDN. You push your changes to GitHub and we take care of the rest with Firebase app hosting.
7:01 · We also launched app hosting to GA at Cloud Next, but there’s so much more that we’ve added. We’ve also launched support for deploying from local using the Firebase CLI, so you don’t have to push your code to GitHub. We also now have Terraform support so that you can deploy with any custom images and we’ll have and that’s especially a big ask for our really big customers. And of course, we also have automatic configuration of the Firebase SDKs using the emulator suite. So now, if you’re using Firebase emulators, app hosting just works out of the box.
7:35 · Firebase app hosting also seamlessly integrates with Firebase Studio, which we’ll talk about in a second, so you can easily deploy your prototypes.
7:44 · The model context protocol, or MCP, is an open standard designed to enhance the capabilities of AI agents. Raise your hand if you’ve used MCP or heard of it.
7:55 · Awesome. About half the room.
7:58 · You can kind of think about MCP like a universal remote or USB for AI applications. It’s like standardizing the way that they connect and inter interact with the outside world and external data and tools and services.
8:10 · Now with MCP support coming from Firebase, you can use it in the CLI to access the full power of the Firebase services with your favorite agencies inside of Firebase Studio 2.
8:22 · We launched Firebase Studio at Cloud Next this year and it should look very familiar to a lot of you. That’s because we were here at Nerdella 2024 and we showed off project idx and this is the evolution into Firebase Studio which is deeply integrated into the Firebase platform.
8:40 · Now let’s talk about green field applications. When we have an idea for an application that solves a specific task or workflow, we need to do two one of two things. One, we can build it ourselves and figure out the framework that we’re going to use, a deployment strategy or services. Or we find an engineer to take time from what they’re doing and build this prototype for us.
9:02 · But now we have a third option, and that’s to be able to generate a prototype from scratch without installing anything.
9:09 · We can launch Firebase Studio and give it an idea for an application using the app prototyping agent.
9:16 · In just a few moments, we will see a blueprint of everything that is in the app that it should build. We can take this time to refine it as needed, provide it more context and about the idea, remove features that and we do not need and after we review and ideulate, we can hit generate. Firebase Studio builds a full stack next.js XJS application that uses Tailwind for CSS and Genkit for AI features.
9:42 · You do not have to install anything since Firebase Studio is built on top of Google Cloud Workstations and Nyx OS. We have a single file called dev.nix that we will provision a machine for you and manage all the dependencies for any framework that you choose. Nyx OS is especially cool because it’s backed by a massive amount of packages over 100 or 120,000 packages for Linux and you can define them in a single file and it’s kind of like Docker but for your operating system.
10:11 · After we see our app running, we can interact it with a built-in preview.
10:16 · This even has the option to share via QR code so that anyone with a link can access it on local host running on the VM.
10:25 · This is not intended for serving, but it makes it easy to get started and get quick feedbacks on prototypes that you’re not ready to deploy yet.
10:35 · With applications that require AI features, Firebase Studio will help you generate an Gemini API key. You don’t even have to leave the console and we’ll take care of everything for you by giving you the correct APIs in the Google Cloud console. And we’ll even create a Firebase project for you if you don’t have one already.
10:52 · But what if we want to connect to a real database or authenticate the user?
10:58 · This is something that people have been asking about for a while. Well, we recently launched support for connecting to Firebase authentication and fire store right in the chat. We can prompt it to add some features that will require a database or O and we’ll provision the resources, set up the rules, and modify any existing screens for you.
11:18 · Firebase makes it easy to move beyond the prototypes and connect to real data sources so that you can move quickly with confidence. We’re working on so many other integrations and we’d love to get your feedback. This is available to everyone now. So definitely go ahead and try to get started.
11:37 · When you’re ready to deploy and share your prototype with the world, you can hit publish and we will roll out the application to Firebase app hosting for you. So anyone with a link can now access the app that you just built.
11:51 · But what if you’re wanting to use a specific stack? Oftent time people come up to us and they’re like, I don’t want to use Nex.js, but I want to use a specific framework that I love or I want to customize it. Well, Firebase Studio has a lot of templates to get started with. And we recently updated over 10 templates in Firebase Studio to make it easier to work with Agent tooling. We added support for the Firebase MCP server, added rules and files, as well as building a new default workspace layout that let you get started building with Gemini right away.
12:21 · The built-in model in Gemini and Firebase now uses Gemini 2.5 Pro, and this is also something people were really excited about. This will greatly improve the quality of the apps being built as well as the issues being debugged.
12:38 · There are many more templates to check out and if we’re missing any, please submit a PR to our open- source templates on GitHub.
12:49 · Now, let’s uh now let’s build an app with the Angular template.
12:54 · This is one that we recently updated and this template will also include the new Angular MCP server and it will improve the the chat experience in Firebase Studio.
13:05 · With the Angular template open, we can start describing what we want to build.
13:09 · Let me see if I can get this started.
13:16 · This demo is not running, but um similar to the prototyper, similar to the prototyper, it’ll come up with a blueprint and a PRD that’ll generate a plan and start building out the application. But we’ll go over a Flutter one in a second.
13:34 · If you prefer the Gemini CLI, which raise your hand if you’ve used the Gemini CLI before.
13:39 · Awesome.
13:41 · It’s an amazing tool and you can open up the terminal and we have it pre-installed for you. All you have to do is log in. The AI rules I mentioned will work right away with Gemini CLI and I personally love using them both together as they can help in different use cases. One of the use cases I like to use Gemini CLI for in Firebase Studio is for search grounding and URL context to research topics and save them as markdown files for context in Gemini and Firebase to use.
14:11 · But what if we need a mobile application and the web will not fit the use case that we’re looking for? Well, Firebase Studio has both Flutter and React Native templates that we can use to build mobile applications without having to set up any SDKs locally.
14:26 · We can go to the recently updated Flutter template and create a new project. This will launch in the same way as the Angular template and we’ll now have access to the Dart MCP server that we can leverage in Gemini and Firebase.
14:39 · This makes tasks like formatting and checking for errors and debugging layout issues a lot easier with Gemini. Because the MCP server is already set up for you, you have the tools ready for you in the chat anytime with Firebase and Flutter MCP ready to go. And just like the Angular template, we can prompt Gemini to build an app and it’ll create all the files that we need. One of the useful tools that I like in the Dart MCP server is searching pub.dev and asking for routing and state management packages. It will search pub.dev and then add them to the project.
15:10 · When the app is done, you will see a real Android emulator running in Firebase Studio as a second preview option.
15:22 · We want to meet you where you are and provide AI assistance, inference or agents wherever you need. Firebase is full stack Google for the AI era and we cannot wait to see what we’re going to build. Thanks.
15:35 · [Applause] Thank you so much. You can stand here.
15:46 · We can make some questions.
15:48 · Sounds good.
15:48 · If you want it.
16:01 · address.
16:11 · No.
16:19 · No worries.
16:28 · Hi everybody. Uh thank you for being here. Um let’s say I’m a boost trouble founder and like alone or a little team. So I want to validate a PC or MVP and let’s say I want to use favor studio.
16:43 · Should I focus my energy uh to in order to do that on which skill would you say like prompt engineer, UX, UX research or should I try learning essentials to become a developer?
16:58 · You know, I would say one of the things that’s really important in this AI era for building um is UX because it’s the intersection between both building and UI. uh if you are generating experiences, they don’t really matter if someone can’t use them. So, making sure that you focus on what is the actual end user uh experiencing as they’re building.
17:19 · I’m not uh I’m not a technical person.
17:23 · So, would you say like uh should I invest my energy in UX entirely?
17:29 · So, UX is in between development and just designing. So it’s when you’re using AI to just build prototypes, you don’t have to be that technical, but rather it’s about understanding the problem and what you’re trying to build.
17:45 · So you’ll work with PMs and teams on seeing what the customer user journeys are. And you’ll work with engineering to see what is trying to be built and what needs to be tested and a and a really good UX can come up with ways that we can improve the experience uh a lot.
18:02 · Okay, got it. Thank you so much, man.
18:04 · Other questions? I want have a question and if is it possible to deploy the front and back end and database like all together with a configuration already like done with a template?
18:23 · Yes.
18:23 · So Firebase Studio like I said has dev.nix and one of the cool things about Nyx OS is it’s a full Linux machine. Uh we have many things ready to get set up such as Docker which you can add with one line. You can use Kubernetes. Um something that I’ve recently explored and is really popular is like caddy uh like caddy files for reverse proxy as well as like traffic. Um all you have to do in Firebase Studio which we have templates for is set up your different services on different ports. Uh and in the sidebar you can even expose those for testing.
18:55 · And when you’re ready to go to production, all you have to do is set up a routing template. So you can say, I want the front end to go to this URL and the back end to go to this URL. And you can deploy it to cloud run or any other places including Firebase app hosting.
19:10 · Okay. Thank you. And like if I have an already like an existed application and I want to integrate with Firebase, how we see it?
19:22 · So uh we also offer a really good import for GitHub. GitHub, Bitbucket, and GitLab. And when you import that, you can create a dev.nix file. Um, if you don’t have Firebase in the source code, but you do have Firebase um, like as a service that you want to set up, you can use Gemini in Firebase to ask it. It’ll have the Firebase MCPS server now, so it can help provision you and walk you through that. If you already have a Firebase project and you import it, all you have to do is Firebase login and Firebase init, and it’ll show an account picker that you can choose your project.
19:55 · Okay, thank you.
19:58 · Hi, how are you? Um, I came here last year and I won a IDX t-shirt. So, it’s quite unique, isn’t it?
20:05 · Yeah.
20:05 · Well, the question is like um the team at Fire Race Studio was all the time making IDX or it was like a side project and that stuff and then merged into Fire Race Studio that now we know they merge. Um, as with many things at Google, you know, uh, sometimes we work on things and we’re not sure how they’ll manifest.
20:28 · And idx was why it was called project idx was a preview that we wanted to build out a technology to get feedback from users. You know, one of the things that idx aimed to solve was enabling computing for many people so that even if you just had a Chromebook or a web browser or a phone, you could still access a full version of VS Code and have all the services like a built-in preview for you. Um, with Firebase, we wanted to and heard from a lot of people that they wanted to have deeper integrations for um, using storage or O connecting to a database.
21:02 · And by having it all under one brand really helped unify the teams and to move quicker.
21:09 · Thank you.
21:10 · Terrific.
21:12 · Also, I understand that Firebase Studio has a lot of integrations.
21:15 · Yes, lots. Yes.
21:17 · So that makes easier to create full applications.
21:20 · Mhm.
21:22 · You have another question there.
21:26 · Hello. I have a question. Erh, what kind of things you can do with Unity?
21:35 · Uh, with Unity.
21:37 · Yes, I see Unity in the presentation.
21:41 · So, Firebase supports Unity in our SDK for Android XR and a bunch of other uh efforts around that. Firebase Studio uh doesn’t support Unity um because it u just runs um uh web applications and desktop. But uh if you’re wanting to use Unity, you can still do it locally.
22:01 · Okay.
22:01 · Uh with this agent, with MCP, you can create anything more different knowations or anything more. Yeah, the MCP server basically enables a bunch of features that you would have to do in the console. So like setting up a database and configuring rules, creating new resources, and now you can just stay in your favorite agent and it’ll do that for you.
22:41 · Fire.
22:49 · Okay. If I raise to your support okay good any other question around hello nice what advice will you give to people to start using fire studio in projects that are start are right now in production or had so many hers um so One of the things that we’re seeing a lot of people build with is uh a couple different things. One, people love to build games in Firebase Studio.
23:26 · So, if you have something that you’ve been wanting to try and create, it’s a great spot to give a first draft. Um, people with a data science background are really liking Firebase Studio because they can upload documents and create these one-off workflow uh tasks.
23:40 · Um, and then also people are using Firebase Studio for like staging environments because you can import uh your code into this VM. You can set up Postgress. You can set up all of your features. And because of our really cool sharing and collaboration features, you can share that with other people at the company, all be connecting to it at the same time. And you can even have your mobile applications running locally connect to the the staging VM.
24:08 · I have another question if you want.
24:10 · Looking at the Firebase road map, what is the next big step of AI power experience for developers? Uh so as we shared earlier Firebase AI logic is something that uh we are investing a lot in and right now we’re looking for the feedback because we are building out this SDK to make as easy as possible to bring it to as many platforms as possible like in the languages you write.
24:33 · Perfect.
24:37 · Nice there.
24:44 · Okay.
24:46 · Hello.
24:48 · Um, I want to know how you can deploy an an application because I know how this works for front end and back end. That’s is clear for me. But sometime when you needs to deploy a application, maybe you need a a Google account, a account. So I don’t know if exists an integration for for the deploy at production for Google Play you mean like Android?
25:20 · Yeah.
25:21 · Um so we do support uh Flutter and React Native. Um outside of like just being able to build the APK and the container and upload it to resources. You can totally sign the app bundle uh all built into the VM. Uh so you wouldn’t have to do anything outside of that other than open up Google Play, open up Firebase Studio and just download and upload the files. Uh and then you can also use a variety of like thirdparty services that connect to GitHub and you can push your changes and automatically deploy through CI pipelines.
25:54 · We have another question uh in the chat. You want to read it?
26:01 · Uh you can go for it. I will get fired.
26:04 · I’m not sure.
26:06 · Being a dev, I use V0 and the versels AI SDK daily. What do you think in this ma is the main advantage of switching to Firebase Studio and Genkit?
26:19 · Um, so first of all, Firebase Studio is a VM that you own. So you can use all the things that you would normally inside of Firebase Studio. So if you connect to Verscell SDKs or you decide to deploy outside of Firebase, we just want to make it as easy as possible to integrate. Um, but one of the things that Genkit offers is the open telemetry to be able to connect to all of the monitoring. Uh, we’ve seen people use it a lot for evals to be able to see in production um, are my prompts working?
26:48 · And that’s where things like Firebase remote config come in really handy. So you can dynamically change the model that you’re using in production. You can change the prompts if you’re seeing that there’s like um starting to see a lot of abuse and you want to lock it down. You can do that without having to deploy an update.
27:03 · Nice.
27:04 · Yeah. And another question, do you have access to any specific models when when you use Firebase Studio? You have different quotas or tiers.
27:12 · Yeah. So, uh everybody for free gets uh three uh workstations and we have a very generous Gemini 2.5 quota u very similar to Gemini CLI.
27:22 · Okay. So if you you Okay. Do you have a question?
27:26 · Uh yeah, here. Um so I’m asking out of ignorance because I’ve never used Firebase before. Um is there a way to change the subdomain once you deploy an application to production? Because I’m guessing that when you deploy you get an URL with Firebase, like a subdomain. Is there a way to bring your own subdomain?
27:46 · Yep, you sure can. Uh, one of the cool things about app hosting and the regular Firebase hosting is not only can you connect any custom domain, you can also do multiple subdomains. So, you can even have different domains for preview ser uh branches or different micro apps if you have like a monolith.
28:02 · Okay, super cool. Thank you.
28:05 · Good.
28:06 · Right.
28:08 · Hi, Rodie. Very nice presentation. My question is not technical. uh the company I work for is pretty much involved or locked in I would say with uh Microsoft solutions and Ashure. What would you say I I can say to them in order to take a look on this or consider this this this solution?
28:28 · Um so one of the cool things about Firebase Studio and um the previous tool idx is we’re just trying to make it as easy as possible for people to onboard.
28:37 · We hope that if you brought in your company’s code and created a template that people could onboard writing code right away within minutes instead of hours or days like it normally would take. Um so right away like people getting into it are not familiar with their codebase can already start to build.
28:52 · Uh the other thing is it’s really good for um ramping up to new products or like if someone wants to integrate a new feature or library or drastically change what’s needed if you’re going from Python 2 to 3 or completely switching to Go and a Rust backend, all they have to do is update the dev.nix file and nobody has to like install anything locally.
29:16 · Asa, hello. Thank you for the presentation.
29:19 · Uh this is regarding MCP. Yes, we support um Firebase Studio.
29:24 · I need to go again. Yeah. Hello. Thank you for the presentation. This is regarding MCP. Uh I don’t want to put you on the spot. Uh but regarding the I just wanted to know which are all the native features supported by the MCP that are allowed at this point. I mean it is only tools. It supports also prompts resources or or the other ones.
29:45 · Yeah.
29:45 · So Firebase Studio uh I believe supports tools. Uh and then Gemini uh CLI inside of Firebase Studio uh supports a lot more. There’s also Genkit in Firebase Studio which also can read and be an MCP server and it also supports uh a lot of the uh the different aspects of MCP2.
30:06 · Okay. Do we have any other question?
30:09 · Guess I have more questions.
30:11 · Yeah, me too.
30:13 · Yeah.
30:13 · For example, the first time I saw Firebase Studio, it was like it’s a VS Code in the cloud, right? But when I saw the demo in the stand yesterday, I saw that you have like a lot of new things, integration stuff. What are like the top three things that you can do in um Firebase Studio that you can do easily on a local development?
30:33 · Yeah, I think one of the things uh like I said that people may not realize is just how easy it is to get a database up and running or get these like really weird obscure uh examples that you find on GitHub. Um you can use Gemini to help you like figure out, oh, this is not installed and it’ll go and search and get it updated for you. Um, we also have some really cool things like the um setting up the Firebase services.
31:00 · Uh, because we’ve done the work to integrate it so deeply into Firebase, uh, we will go ahead and do like 10 steps that you would have to do by hand all just behind the scenes like clicking for you. So, if you’re just in that universe and you want to use us, uh, we were going to try to make it as easy as possible. Um, and then even simple things like creating a Gemini API key usually takes like a couple clicks or if you’re having to go into AI studio and grab it from different places, we keep it all in one place uh for you ready to go.
31:30 · Nice. And when you are using the MCP server for Firebase, you can also query information about your database inside. So something I usually do is go to the Firebase um like dashboard and go to the database try to query.
31:45 · You can just ask it in the chat now.
31:46 · Yeah.
31:47 · Yeah.
31:47 · how many users are in my application and you will get yeah that’s quite nice perfect any other question I have the last one if you want it oh sorry hello my name is Timur I’m a data engineer and primarily firebase is a as a tool for software engineer what about uh Data engineering.
32:18 · Can we data engineers leverage uh this tool for our daily tasks?
32:24 · Oh yeah. Uh one of the cool things about Firebase Studio specifically is like you don’t even have to use Firebase if you don’t want to. You can open it up and create like a data lakeink with Duct DB or connect to BigQuery. Um a lot of people use it to throw in a bunch of different connections and create little dashboards to be able to visualize the data. So yeah, that’s absolutely something that we saw really early.
32:45 · People are excited about it.
32:48 · Nice. And the last one if you want it.
32:51 · Sounds good.
32:51 · What is the AI power experience trend that you are most excited about right now? There are any application or features being built with Firebase that you consider revolutionary.
33:04 · I still feel like uh Gemini Nano Banana is still really popular and people building some awesome stuff with it. Yeah. And I think uh that’s been an explosion of like creativity that has been really cool to see especially when it comes to AI logic or genkit. People are just asking to generate images and that seems to be something that a lot of people want to do in our products.
33:27 · I don’t know if people try raise your hand if you try nanoano.
33:33 · Nice.
33:33 · We need to see a lot more hands for sure.
33:36 · Yeah.
33:36 · No, it it’s really fast. I I think it’s the best thing it has compared to other models. you have like flu flux context that that it’s really nice but it’s not as fast as another one.
33:46 · Yeah. And it’s also really easy to edit images and other things too.
33:49 · Yeah.
33:50 · And the is so good.
33:51 · Yeah. Yeah. Yeah.
33:52 · Really?
33:53 · Yeah.
33:53 · Awesome. Well, [Applause] thank you. One more time. Thank you.