If we talk about personal assistants, the first thing that comes to mind is the one we can find on our PC or mobile phone, be it Siri, Bixby, Cortana, Alexa or the Google assistant (which does not have its own name). These assistants are responsible for searching for information on the Internet to answer our questions in addition to being able to perform certain actions related to the device where it is executed.
However, as we always say, imagination has no limits and if not, tell Manuel Ahumada who has created his own virtual assistant based on Benderthe legendary beam-bending, cigar-smoking robot from the series Futurama.
When we say based on the character from the Futurama series, we mean that it has used both the voice (from the American version dubbed by John DiMaggio) and the design of the head. To carry out this project, much more complete than any other personal assistant, he has used, of course, a Raspberry Pi and ChatGPT in addition to other AI platforms and, of course, a 3D printer to create the case.
A unique personal assistant
Ahumada has built not only a unique and very original personal assistant, but it is also much smarter than the rest of the personal assistants that we can find on the market (Siri has always made it very easy).
Interaction with personal assistants is limited to asking questions about any issue. Ahumada has gone much further since it also incorporates facial recognition to follow the person you are interacting with.
For this assistant to be able to interact visually and through voice commands, Bender sends the questions to ChatGPT to obtain a response in text format, response that is passed through Eleven Labs to convert the text into a voice response with the voice, pardon the redundancy, of John DiMaggio, who, as we mentioned above, is the person who dubbed Bender in the United States.
The use of the camera is not limited to following the user, but can also be used to show objects and even videos so that it offers additional information with its characteristic sarcastic tone. To carry out this functionality, it uses OpenCV.
As we can see in the video above, by activating the recognition function, it is capable of detecting any type of content and offering information about it, be it books or even retro video games like Altered Beast which Ahumada plays during the video.
Inside, there is a Raspberry Pi 5 Connected to a microphone, a speaker, the camera and several motors are responsible for moving the eyes to give a greater sensation of realism. Unlike other projects created using a Raspberry Pi, the only information available to carry out this project is the same that we have shared in the article and that is also available in the video published on YouTube that you can find above these lines.
You know, if you want create your own personal assistant, you only need a Raspberry Pi, ChatGPT to generate the responses, OpenCV to recognize and describe objects and Eleven Labs, so that your assistant has the voice of the celebrity you want. We cannot forget a 3D printer to be able to shape it.
