I built BillAI Bass just because I could. It’s powered by a Raspberry Pi and Strands Agents with bidirectional streaming, and the repo includes the code and hardware setup. Happy to answer questions about the implementation!
In lieu of the current lip-sync mechanism, do you have any plans to try giving the AI model more verbs to work with the motors (within safe thresholds), and then encouraging it to do more than just lip-sync? Lip-sync should clearly be the default for normal verbal responses, but you could also encourage responses like jaw-drop reactions, or patterns (would be funny to see it using its tail to tap out "shave and a haircut" as a bit).
If there is a small model that can run on a low powered device you might be able to use that too so that it is all self contained.
It will think (generate tokens) throughout the day, then convert that to voice and save it and will keep generating so on. Every time it is poked, it will play one of the saved recordings.
Someone else suggested Gemma E2B which I can run on Ollama def going to try that next. I can also write a lil cronjob to have it wake up and do things on its own then wire up the button to kick off a function that starts the voice mode. I’ve found local models sometimes struggle with tool usage though. Fun direction to try next. Thanks!
That’s awesome! Billy bass has always been a fave for this sort of thing I found a couple of guides that were similar to mine when I was working on it too
+10 points for giving it context that it's nailed to a plaque. -10 points for using Amazon Bedrock instead of a local conversational model that can run on the Pi without a subscription fee.
You can actually run a local model with strands using Ollama pretty easily so you could just swap the model provider out in like 4 lines of code. So this would not be hard to do. Good idea!
Yea, I like the idea of hooking this up to a local model running on my homelab and having no external calls. This project seems to provide a pretty good hackable foundation! I also wonder if it can be done with something smaller form-factor like a raspberry pi zero so it can be fully encased in the Bass plaque. This is a site to bookmark for later!
EDIT: Looks like modern versions of Billy have a motion sensor too--that could be fun to hack as well.
I’ve seen other people hack Billy bass using other tech but they 3D printed a new casing for it. I looked for it again but couldn’t find it to send you a link though.
This is great! I remember the commercials for these things as a kid. Imagine telling that kid that in the future Billy Bass will have the world's collective knowledge and you'll be able to talk to him about anything.
Yess a nice dose of nostalgia. I want to download the original songs so he can still sing those too and just load them up as audio files it can play. I had it reading the news and giving me a summary but it just ultimately was too silly for that use case lol
The trick for better quality low-skill soldering is:
1) Use a leaded solder with flux mixed in like the Multicore 502 [0]
2) Use more flux.
3) Use little alligator clips on adjustable arms to hold the stuff together for you so you can keep adding… yep… more flux.
4) Get some braided copper solder wick with flux in it for when you’ve finally added too much solder but it still looks terrible and you just wanna start over. After wicking away all the solder, clean up the wires with some more flux.
Oh also the flux is more toxic than the lead. Get some good ventilation so you can use more flux.
I built BillAI Bass just because I could. It’s powered by a Raspberry Pi and Strands Agents with bidirectional streaming, and the repo includes the code and hardware setup. Happy to answer questions about the implementation!
Fun project!
In lieu of the current lip-sync mechanism, do you have any plans to try giving the AI model more verbs to work with the motors (within safe thresholds), and then encouraging it to do more than just lip-sync? Lip-sync should clearly be the default for normal verbal responses, but you could also encourage responses like jaw-drop reactions, or patterns (would be funny to see it using its tail to tap out "shave and a haircut" as a bit).
lol a jaw drop reaction could be awesome. Or like maybe extra tail flaps when it thinks it’s warranted.
If there is a small model that can run on a low powered device you might be able to use that too so that it is all self contained.
It will think (generate tokens) throughout the day, then convert that to voice and save it and will keep generating so on. Every time it is poked, it will play one of the saved recordings.
Someone else suggested Gemma E2B which I can run on Ollama def going to try that next. I can also write a lil cronjob to have it wake up and do things on its own then wire up the button to kick off a function that starts the voice mode. I’ve found local models sometimes struggle with tool usage though. Fun direction to try next. Thanks!
Reminds me of Mike Neil's FishHack at the 2000 MacHack. https://tidbits.com/2000/07/03/the-machax-best-hack-contest-...
That’s awesome! Billy bass has always been a fave for this sort of thing I found a couple of guides that were similar to mine when I was working on it too
+10 points for giving it context that it's nailed to a plaque. -10 points for using Amazon Bedrock instead of a local conversational model that can run on the Pi without a subscription fee.
Replace the bedrock setup for a gemma-4 E2B one.
You can actually run a local model with strands using Ollama pretty easily so you could just swap the model provider out in like 4 lines of code. So this would not be hard to do. Good idea!
Yea, I like the idea of hooking this up to a local model running on my homelab and having no external calls. This project seems to provide a pretty good hackable foundation! I also wonder if it can be done with something smaller form-factor like a raspberry pi zero so it can be fully encased in the Bass plaque. This is a site to bookmark for later!
EDIT: Looks like modern versions of Billy have a motion sensor too--that could be fun to hack as well.
I’ve seen other people hack Billy bass using other tech but they 3D printed a new casing for it. I looked for it again but couldn’t find it to send you a link though.
This is great! I remember the commercials for these things as a kid. Imagine telling that kid that in the future Billy Bass will have the world's collective knowledge and you'll be able to talk to him about anything.
Yess a nice dose of nostalgia. I want to download the original songs so he can still sing those too and just load them up as audio files it can play. I had it reading the news and giving me a summary but it just ultimately was too silly for that use case lol
> was too silly for that use case lol
I don't know, some of the news today is so preposterous that it could only be delivered by a BillyBass.
I love that you put the Amazon Shopping list in the readme.md. Hacking this together is not intuitive
Yes I agree! not intuitive at all. I tried to lay out everything I did so someone could easily recreate it.
we need more people like you. kudos
<3
What a great name for a project
:)
Get a proper wire stripper! Don't strip wires with a diagonal cutter.
Hahaha it was my first time!!! I am but a mere software person. You wouldn’t want to see my soldering joints… it wasn’t pretty but it got the job done
The trick for better quality low-skill soldering is:
1) Use a leaded solder with flux mixed in like the Multicore 502 [0]
2) Use more flux.
3) Use little alligator clips on adjustable arms to hold the stuff together for you so you can keep adding… yep… more flux.
4) Get some braided copper solder wick with flux in it for when you’ve finally added too much solder but it still looks terrible and you just wanna start over. After wicking away all the solder, clean up the wires with some more flux.
Oh also the flux is more toxic than the lead. Get some good ventilation so you can use more flux.
0: https://www.newark.com/multicore-loctite/c-502-99c-5c-0-7mm-...
Okay so what I’m hearing is more flux. Thank you for these tips I’ll look into it!