Our ideas are shared through the wondrously flexible medium of storytelling and now, thanks to maker and developer Stefano Mazzocchi, you can take this delightful human experience and make a Raspberry Pi do it for you. Mazzocchi has created a device called Fably that uses the power of AI to generate stories. Fably has been made open source and is designed with a user-friendly interface that even children can use.
Fably is built upon AI-focused cloud-based APIs, leveraging these systems to create bedtime stories. The device also doubles as a platform you can use to play pre-recorded audio books that can be much more immersive and lifelike than listening to text-to-speech voice-generated audio. At the end of the day, it's your call to decide if you want to delve into a story created by a human or not.
According to Mazzocchi, another goal was to keep the project accessible by designing it to operate using less than $50 worth of hardware. However, it's worth noting that one of the APIs used in the project operates with a contract to charge you for each request it processes. You can see a video of Fably in action over Mazzocchi's YouTube channel.
Mazzocchi demonstrates Fably's ability to run on a Raspberry Pi Zero 2 W, which costs around $15, but you'll need a few extra components like a speaker, microphone, power supply, housing, etc. If you don't have a Raspberry Pi, Mazzocchi has also made a version available that works on Windows and MacOS machines, so you can check it out for yourself.
Fably accepts voice commands for input, so you don't need a keyboard or any buttons. Mazzocchi also mentions looking into Ollama, a language model that Fably is compatible with. The project is very scalable to suit a variety of hardware. You can use what you've got on hand, or get something specific that caters to the project's needs. It would also be beneficial to contain the hardware so that it's safe for children to interact with.
If you want to get a closer look at this Raspberry Pi project in action, you can check it out over at Hackster and peruse the source code at GitHub.
