Raspberry Pi-Based DIY Robot 'Saras' Combines Speech, Vision, and Autonomy

Published  July 9, 2025   0
User Avatar Abhishek
Author
Saras: Raspberry Pi Robot With Autonomy

Creator Jack Of All Tech built a personal robot that can move, speak, listen, respond, and make decisions all on its own. It’s called Saras, and the name stands for Smart Autonomous Robotic AI System. It’s made up of a Raspberry Pi, some sensors, wiring, 3D printed body parts, a microphone, a speaker, and DC motors. It has a wake word system, so it does not react to random sounds, just like what we see in Alexa or Google Assistant. The phrase "Hey, Saras" acts as the trigger.

The screen is simple; it shows faces to indicate its current status. Blinking means listening, talking face means speaking, and neutral face means idle. The robot can listen and use AI models to formulate responses. It turns speech into text, sends it to ChatGPT, the local LLaMA model, or the local LaVA vision model, then speaks the answer using text-to-speech. The robot can move forward, backward, left, or right according to verbal commands thrown at it. It can basically see a box or a couch and move towards it on command. It interprets the command, identifies the object, and acts without bumping into anything, demonstrating obstacle detection and spatial reasoning.

It has a full autonomy mode where it explores an area on its own and presents a summary afterwards. The storage of the navigation history improves decisions. Every move is logged: the direction, distance from obstacles, and what’s seen. These logs can be used to build a 2D map of the house. Future improvements could include a more polished body and an offline mode. 

Add New Comment

Login to Comment Sign in with Google Log in with Facebook Sign in with GitHub