Building IoT with Serverless AWS: A Cloudagotchi Playbook

A founder built a connected pet, Cloudagotchi, with its core logic and state residing in AWS. This project demonstrates a serverless architecture for IoT, shifting device intelligence to the cloud.

A founder on dev.to reports building a connected device, Cloudagotchi, with its operational brain entirely in AWS for "well under a dollar a month" in running costs. This project details a serverless architecture for IoT, shifting device state and logic from embedded hardware to the cloud. The approach decouples the physical device from its core functionality, allowing for persistent state and complex interactions without constant device uptime.

Cloud-Native State Management

The Cloudagotchi's core state — its hunger, emotions, and interactions — resides in AWS, not on the physical ESP32-S3 board. The founder explains that the pet "gets hungry while the device is off" because time passes in the cloud. EventBridge Scheduler is used to trigger events like dinner time, ensuring the pet's needs evolve independently of the device's power status. Its feelings are stored as a DynamoDB row, preventing a simple reboot from resetting the pet's emotional state or past neglect. This architecture ensures continuity and persistence, a critical component for any connected device where user interaction might be intermittent.

MQTT for Device Interaction

User interactions with the Cloudagotchi, such as tapping an avocado to feed it or shaking the board to play, are translated into MQTT messages. These messages are sent via mutual-TLS MQTT to AWS IoT Core. This service acts as the central hub, securely ingesting data from the device and routing it to the appropriate serverless functions. The use of mutual TLS provides strong authentication, ensuring only authorized devices can communicate with the cloud brain. This pattern is standard for robust IoT deployments, providing a scalable and secure communication channel between edge devices and cloud services.

AI-Powered Personalization and Output

Beyond basic pet care, the Cloudagotchi integrates advanced AWS services for unique features. Every morning at 7:03, the pet reads the AWS news aloud. This is achieved through a pipeline involving Amazon Bedrock, which rewrites headlines in the pet's "cute voice," and Amazon Polly, which converts the text into speech. The audio is then streamed, and the device's face animates to simulate talking. This demonstrates how serverless AI/ML services can add personalized, dynamic capabilities to connected devices, moving beyond simple sensor data or command execution.

Infrastructure as Code with CDK

The entire AWS infrastructure for Cloudagotchi, encompassing IoT Core, Lambda, DynamoDB, EventBridge Scheduler, Bedrock, Polly, and S3, is defined using AWS CDK. This infrastructure-as-code approach ensures that the cloud brain is reproducible, version-controlled, and easily deployable. For indie founders, CDK (or similar tools like Terraform) provides a structured way to manage cloud resources, reducing manual configuration errors and accelerating development cycles for complex serverless architectures.

What We'd Change

While the Cloudagotchi project effectively showcases serverless IoT capabilities, its direct applicability to a commercial product requires modification. The reported running cost of "well under a dollar a month" is for a single, personal device. Scaling this architecture to thousands or millions of users would introduce different cost dynamics, particularly for DynamoDB read/write units and IoT Core message throughput. A commercial product would need a more granular cost analysis and potential optimizations for multi-tenancy.

The custom UI built with LVGL on a specific Waveshare ESP32-S3 board implies a tightly coupled hardware-software solution. Commercial IoT products often prioritize hardware abstraction or broader platform compatibility to reduce reliance on single-source components and simplify manufacturing. Furthermore, the initial blog post mentions an "IoT policy trick that makes a kidnapped pet a non-event" but does not detail its implementation. Robust security measures, including device provisioning, credential rotation, and secure firmware updates, are paramount for any commercially viable connected device and would require explicit, verifiable strategies.

This Cloudagotchi project offers a concrete, open-source blueprint for indie founders exploring connected devices. By offloading complex logic and state to serverless cloud infrastructure, it simplifies embedded development and enables rich, AI-driven features. The approach demonstrates a viable path for creating sophisticated IoT products without the overhead of maintaining dedicated servers, making niche hardware-software integrations more accessible to small teams.

The investor read

The Cloudagotchi project highlights a growing trend in connected device development: leveraging serverless cloud infrastructure to reduce hardware complexity and accelerate feature development. This approach lowers the barrier to entry for indie founders in the IoT space, enabling the creation of niche, intelligent devices without significant upfront infrastructure investment. The use of services like Bedrock and Polly signals a shift towards AI-enhanced, personalized device experiences. While the reported sub-dollar monthly cost is compelling for a personal project, commercial scalability would require careful consideration of AWS service pricing at volume. Investable propositions in this segment would need to demonstrate robust security, multi-tenancy capabilities, and a clear path to market beyond a single-user prototype.