Delivered in-demand app capabilities that sparked a 30% to 40% engagement rate
Reduced image generation latency to seconds for faster user experiences
Built and delivered creative new app features in less than a week
Gave users the ability to customize photos with simple text prompts
Remixing everyday selfies for your inner circle
Your phone’s home screen flashes to life with a picture of your best friend laughing halfway across the country. That instant spark of personal connection is at the core of Locket’s success. What started as the founder’s gift to his girlfriend graduating from college quickly grew into a global community seeking a more personal alternative to traditional social media. Gen Z and Gen Alpha embraced the platform because it focuses entirely on fostering close connections, not building an impersonal community of followers.
In just a few years, Locket grew to serve over 10 million daily active users, with Millennials and Gen X increasingly downloading the app. Today, the company continues to evolve user experiences, most recently exploring how AI can further enhance engagement while staying true to Locket's commitment to enabling authentic, real connections.
The product team envisioned an interactive feature called “Locket Looks,” where people can use AI to add playful overlays to selfies. That might look like a colorful costume or a silly hairstyle—all while keeping the underlying image fully intact and readily recognizable, something most AI image models struggle to do: they often change faces and shapes or erase backgrounds, pulling users out of the moment they want to share.
Additional challenges with AI tools that Locket considered were the high costs and long loading times to update a photo. Locket knew the experience had to feel quick and easy for users to embrace it.
We run a lot of our business on Google Cloud, including Gemini models for advanced safety and security. We realized we could go further with Gemini to enable people to express themselves more creatively. With Google Cloud, we build tools that keep experiences fast, playful, and secure yet still grounded in making authentic connections.
Matt Moss
Founder and CEO, Locket Labs
Delivering lightning-fast, secure photo sharing for millions
From day one, Locket used Google Cloud to fulfill its goal of deepening connections between loved ones. The app was created with Firebase and scaled to millions of users through a full suite of technologies. Locket handles the huge volume of photo uploads by pairing Firestore for instant account management with Google Cloud Storage. Bigtable processes heavy database workloads. For lightning-fast insights, BigQuery handles data analysis in seconds. The fully managed setup built on Google Cloud enables a team of just 15 employees to move from an idea to a live prototype in less than a week.
By running our entire infrastructure on Google Cloud, our lean team can launch and manage incredible features with just a few lines of code. For our new image capabilities, Nano Banana 2 Lite gives us world-class quality alongside the speed and affordability required to serve millions of users.
Matt Moss
Founder and CEO, Locket Labs
Engineers turned to Google Nano Banana 2 Lite when launching the new Locket Looks feature. Locket was a member of the selective early access program to test the new model from Google DeepMind. With just ten lines of code, they unlocked a high-speed pipeline that lets users snap a selfie and type simple text prompts to add creative filters, like a wizard robe, a cowboy hat, or a clown nose. The technology preserves facial traits, expressions, skin tones, and backgrounds, so images stay rooted in reality. Unlike other AI tools that can take minutes to process an image, Nano Banana 2 Lite does it in just a few seconds.
Because user safety on the app is non-negotiable, Locket deployed advanced Gemini models to proactively protect user experiences.
Every photo sent through the app passes through an automated filtration layer to stop explicit content before reaching its target. Gemini also reviews text chats to flag indicators of bullying or self-harm. When needed, the app automatically shares wellness resources to protect its community in real time.
Achieving rapid success while deepening personal connections
The update is an instant hit. Most apps only see about 5% to 10% of their users try a new feature. Locket saw an impressive 30% to 40% of its daily active users jump in right away. This success proved that high-quality AI can be accessible to millions without complexity and high costs.
With Google Cloud supporting its vision, Locket generated 1 million photos within just two days of making the tools more widely available. Because the new filters run on simple text prompts instead of complex custom code, the team can brainstorm fun new visual ideas and instantly turn them into live features. This flexibility is paving the way for potential Locket partnerships with fashion brands, celebrities, and movie studios.
We are showing that AI can be used to create authentic digital experiences that foster joy. Google Cloud is our foundation to safely bring creative expression to our community while maintaining the trust at the heart of our brand.
Matt Moss
Founder and CEO, Locket Labs
Locket Labs is a social platform that connects close circles of friends through an interactive home screen widget.
Industry: Technology
Location: San Francisco, California
Products: Firebase, Firestore, Google Cloud Storage, Bigtable, BigQuery, Spanner, Nano Banana 2, Gemini Enterprise Agent Platform
