Your AI assistant does the thinking. The robot does the walking, looking, listening and talking. A small bridge program on the robot joins the two together, and our one-prompt setup helps you build it.
Pick one of our four bots. The kits take an afternoon or two to build; the Rider-Pi is mostly built (check the Amazon listing).
Pop it on your home Wi-Fi, then connect it to your AI assistant with a small bridge program on the robot's Raspberry Pi. Our setup prompt guides you, but it takes some technical skill.
Once the bridge is working, say hello. It hears you, thinks it over (expect about 2-5 seconds' delay with a direct AI API, or longer through an assistant), then moves and replies in the voice you've picked.
Copy this into your AI assistant, swap in your robot's name, and it should walk you through setup one step at a time, safety limits included. It takes some technical skill and a bit of patience, and it isn't guaranteed to work with every robot or AI.
Hi! I've just bought a home robot and I'd like your help setting it up so I can talk to you through it. My robot: [ROBOT MODEL] (Freenove Robot Dog Kit FNK0050 / SunFounder PiDog / Yahboom Rider-Pi CM4 / Yahboom Rider-Pi CM5) My computer: [Windows / Mac / Linux] My experience: [total beginner / done a bit of coding] Please guide me step by step. Give me ONE step at a time, tell me exactly what to type or click, and wait for me to say "done" (or paste any error) before moving on. Part 1: Get the robot running 1. Prepare the robot's software: - Yahboom Rider-Pi (CM4 or CM5): do NOT flash Raspberry Pi OS. Keep the factory SD card image, because it has Yahboom's drivers and the xgolib library. Instead, help me switch off Yahboom's start-up program so our bridge can control the robot. - Kits (Freenove or SunFounder): install the operating system version the maker's documentation specifies, using Raspberry Pi Imager. Set the hostname, my Wi-Fi, a username and a strong password, and enable SSH. 2. Boot the robot, find it on my network and connect over SSH. 3. Install or update the maker's official Python library and examples for my exact robot model, following the maker's own documentation. If you're unsure of a command, tell me to check their docs rather than guessing. 4. Run calibration if my robot needs it. 5. Test each part separately, one at a time: movement (slowly, on the floor, in one closed room), camera (take one test photo), microphone (record 3 seconds) and speaker (play a test sound). Part 2: Build a small, secure bridge on the Pi Help me write a small Python program on the Pi that lets me talk to you (or an AI model I connect by API) through the robot: - Security first: no open ports on my home router and no port forwarding. The Pi should only make OUTBOUND connections, for example an outbound websocket/HTTPS connection or a private network such as Tailscale. Keep keys in a separate config file that never gets shared or uploaded, and explain each choice in plain English. - Put a secret token on the bridge so only my AI can send it commands, and reject anything without it. - Don't run the bridge as root. Use a normal user account. - Voice in: wake word or push-to-talk, then speech-to-text. - Voice out: text-to-speech through the robot's speaker, with a voice I can choose. - A short, fixed list of commands the robot will accept: move forward/back, turn left/right, look around the room it's in, take a photo and speak. Anything not on the list gets ignored. - Limit and check every movement value (speed, distance, angle and duration) before it reaches the motors, and reject anything out of range. Safety limits (please build these in from the start) - Set a low speed limit and short move durations by default, and keep them low. - Handle "stop" and the button on the robot itself locally on the Pi, not through the cloud, so movement halts straight away even if the internet drops. - Auto-stop if the robot loses its connection or misses a heartbeat. - These robots can't detect stairs or edges. Physically block them off and keep the robot in one closed room. Remind me of this before the first movement test. - If you're unsure what I mean or what's in front of the robot, stop and ask rather than guess. - The camera and mic are only used when I ask, never recording in the background. - A privacy indicator (an LED, the screen or a sound) whenever the camera or mic is on. - A battery-low warning and a safe stop. How I'd like you to work - Never paste passwords or API keys into the chat. Tell me where to put them on the Pi instead. - Test one small piece at a time and confirm it works before adding the next. - Ask me before anything risky or hard to undo (deleting files, changing network or router settings, running things as root, or anything that costs money). - Be honest about what my robot can't do and about any lag. - If something's beyond my robot's hardware, say so and suggest a simpler option. Let's start with Part 1, step 1.
Three parts work together. None of it is magic, and you stay in charge of every bit.
It understands what you say, decides what to do and talks back. It's an AI assistant or AI model you can connect by API.
A small program on the robot's Raspberry Pi. It passes your assistant's instructions to the motors, camera, mic and speaker, and sends back what the robot sees and hears.
Legs or wheels to move, a camera to look, and (on most bots) a mic and speaker to chat. It only does what the bridge tells it to.
The thinking happens online, so each request takes a trip from the robot to your AI assistant and back. Expect about 2-5 seconds' delay with a direct AI API, or longer through an assistant, between speaking and the robot replying or moving. That's normal, and it's why our setup keeps moves short and slow.
Images, audio or text the robot sends to an AI service are handled by that service under its own terms and privacy policy, not ours. Homebody Bots never receives them. Our setup prompt asks for the camera and mic to be used only when you ask. More on that on our peace of mind page.
Check the Wi-Fi name and password (they're case sensitive). Many small boards only like 2.4 GHz networks, so make sure that band is switched on, and move the robot closer to your router for the first setup.
Test the camera in the maker's own app or example first. If that works, ask your AI assistant to check the bridge step that handles photos; paste in any error messages and it'll help you fix them.
Check the bridge program is running and the robot is online. Try a simple command such as "stand up" or "say hello" before anything fancy.
About 2-5 seconds is normal with a direct AI API, or longer through an assistant. Much longer usually means weak Wi-Fi or a busy connection, so move the robot or your router a bit closer together.
Usually it's the battery. Charge it fully and try again. If it keeps happening, check the maker's guide for calibration or servo settings.
Still stuck? Paste the error into your AI assistant along with our setup prompt, or read the FAQ.
This is a general-purpose prompt, not an official integration with any AI assistant or robot maker. Always follow the maker's own safety instructions.
