Cuts infrastructure overhead by 45% to fund innovation with Cloud Run
Reduces latency by 30-70% for real-time app generation using Vertex AI
Scales to 13B+ tokens with zero downtime during global events using GKE
Accelerates feature delivery 2x with unified DevOps tools including Cloud Build
Enabling 140,000+ users to turn natural language into working apps with sub-second latency.
Enabling 140,000+ users to turn natural language into working apps with sub-second latency.
Breaking the barriers to real-time AI creation
ChatAndBuild empowers anyone to build software by simply describing it. It turns natural language into real-time applications, allowing a 4-year-old to design a Peppa Pig math game and a 78-year-old to create strategy simulations. The team believes we're moving toward the "Intelligence-Ownership Economy," where users own Non-Fungible Agents (NFAs): autonomous AI entities that retain their long-term memory, skills, and personality across different models and platforms.
But this vision was hindered by the company’s previous infrastructure. Builders faced 30–70% higher latency, disrupting the flow of real-time generation. Complex agents, such as Japanese language tutors, suffered from context rot, causing loss of user history and hallucinating answers. Global hackathons required engineers to manually provision servers days in advance to prevent crashes from IP surges.
To break free from these constraints, ChatAndBuild sought a proactive partner with an AI-native ecosystem capable of multi-region scaling.
When you’re building a platform to facilitate real-time AI creation, reliability isn’t a ‘nice to have.’ Our previous setup meant engineers were acting as the autoscaler to manually provision, restart, and monitor our systems. Any wobble showed up as higher latency or broken agent sessions, which undermines trust in long-lived agents and memory continuity.
Christel Buchanan
Founder and CEO, ChatAndBuild
Architecting a reliable 'idea-to-app' engine
To replace the stress of manual scaling with an architecture capable of "vibe coding" at global speed, ChatAndBuild migrated to Google Cloud. The goal was to establish an AI-native ecosystem where compute, data, and models operated as a single, orchestrated unit.
Now, the platform handles requests as diverse as a PE teacher building a volleyball velocity tracker or a Head of Product prototyping a customer feedback app. When a prompt arrives, Vertex AI instantly routes tasks to specialized agents for coding and research. For complex multimodal needs, such as analyzing video motion, Gemini models on Vertex AI generate high-fidelity outputs in real time.
Beneath this workflow, Google Kubernetes Engine (GKE) creates isolated namespaces for each agent graph, ensuring thousands of builders can create simultaneously. Cloud Run and Cloud Functions automatically spin up stateless microservices to handle traffic bursts, while Cloud Load Balancing ensures requests reach the nearest region for sub-second response times.
To maintain the intelligence of these agents, the platform uses BigQuery as a telemetry layer, tracking billions of tokens to prevent context rot, and Cloud SQL for structured data management.
Firebase and Firestore synchronize user state in real time, while Cloud Storage and Cloud CDN deliver the final generated assets globally. Supporting this velocity, the engineering team uses Cloud Build and Artifact Registry to ship feature updates twice as fast.
With Google Compute Engine, we can elastically scale our worker fleet from ~10 to 100+ in minutes to support peak demand during a global hackathon. Scaling is handled through instance groups/autoscaling policies, so capacity ramps up automatically based on load and then scales back down as traffic normalises, keeping latency stable while avoiding idle compute. That operational efficiency has been a meaningful breakthrough for us.
Christel Buchanan
Founder and CEO, ChatAndBuild
Achieving global scale without operational stress
The new architecture made the war room anxiety of global hackathons vanish. Instead of manually provisioning servers, the team watched Google Kubernetes Engine (GKE) and Cloud Run auto-scale to process more than 13 billion tokens with zero downtime.
This reliability and the Google Cloud transparent billing policy reduced infrastructure overhead by 45%. This allowed the team to shift their focus from debugging servers to shipping new products such as the upcoming iOS and Android apps.
Google Cloud also unlocked regulated markets. By securing agent deployments with VPC Service Controls, Google Cloud Armor, and Identity and Access Management (IAM), ChatAndBuild can now enable the strict audit logging and SOC2 compliance required by government and enterprise partners.
With a stable, secure foundation, the company is now fully focused on leveraging this AI-native stack to make ownable, evolving NFAs a reality for builders worldwide.
For an AI-native product like ChatAndBuild, Google Cloud gives us the primitives we need to operate at global scale through elastic compute, multi-region reliability, and production-grade security controls. We depend on that foundation to keep latency predictable and sessions stable across millions of real-time agent interactions.
Christel Buchanan
Founder and CEO, ChatAndBuild
Pioneering the "Intelligence-Ownership Economy," ChatAndBuild is an AI-native platform that lets anyone turn natural language into working apps, games, and agents instantly.
Industries: Technology, Startup
Location: Singapore
Products: Vertex AI, Gemini models on Vertex AI, Google Kubernetes Engine (GKE), Cloud Run, BigQuery, Firebase, Cloud Storage, Cloud SQL, Firestore, Cloud Functions, Compute Engine, Cloud Load Balancing, Cloud CDN, Cloud Build, Artifact Registry, VPC Service Controls, Google Cloud Armor, Identity and Access Management (IAM)
