Jump to content

Willow, an open source Alexa replacement that respects your privacy

Summary

  • The good: voice detection is competitive with voice assistants from Nest and Amazon
  • The bad: still a rough experience, DIY, lots of steps
  • Only costs $35 to $50 depending on the ESP-32 box chosen but also needs the backend "inference server" code running on a PC somewhere in your home

 

Quotes

Quote

Willow... utilize a specific set of hardware: ESP-32-BOX devices, which provide a basic hardware shell around ESP32 SoCs. Once flashed with Willow, they can work as locally controlled voice assistants and, eventually, great and easy access points for Home Assistant.

 

In a "Show Hacker News" thread, Willow creator Kristian Kielhofner writes that he wanted to jump-start Home Assistant's voice ambitions by providing ready-to-go hardware. He lays out his pitch for helping develop his first ("and VERY early") release: It's cheap ($50 or less), easily flashed, self-hosted or entirely on-device, open source, and extensible by software or GPIO pins. The voice detection is accurate and reliable, and Kielhofner put up a demo video.

 

My thoughts

It can't play youtube videos yet (or even just stream the audio). But once that happens Amazon Prime music gets crossed off.

 

Sources

https://arstechnica.com/gadgets/2023/05/willow-is-a-faster-self-hosted-diy-voice-assistant-built-on-50-gadgets/

 

Link to comment
Share on other sites

Link to post
Share on other sites

I don't personally see how this is news. It's another open-source voice Assistant. Considering it's self hosted the only thing it really competes with is MyCroft.

If there isn't a big name behind it, it's not in store shelves, and it's not supported by the majority of products then it doesn't matter for the vast majority of people.

Edited by Nayr438
Link to comment
Share on other sites

Link to post
Share on other sites

4 minutes ago, Nayr438 said:

If there isn't a big name behind it, it's not in store shelves, and it's not supported by the majority of products then it doesn't matter for the vast majority of people.

That's probably a bonus for now, once this project becomes more well-known it might attract the people who want something that "just works." I don't see why anyone would have a problem using HomeAssistant to set this up, and the ESP-32 box is already on store shelves.

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, sounds said:

That's probably a bonus for now, once this project becomes more well-known it might attract the people who want something that "just works." I don't see why anyone would have a problem using HomeAssistant to set this up, and the ESP-32 box is already on store shelves.

Hi, project creator here.

 

This is the thinking exactly. Willow is so new, and moving so fast, if we provided an extremely easy way to flash/install/use now it would be disastrous. The users would end up frustrated. They would end up thinking they wasted their money. They'd call it all a scam or who knows what. Perhaps worst of all, as the saying goes "you only get one chance to make a first impression".

We are catering toward users (pretty much developers, really) who look at our build instructions and think "Oh that's not bad". The build process is almost deliberately intimidating as a kind of filter to scare away those who aren't prepared or able to build, flash, and contribute. I don't want people going out and spending money on hardware that they won't be happy with because we're not quite ready for them.

 

This post from HN kind of "broke out" by making it here, to Ars, etc. There's a lot of excitement going around but unfortunately (though not surprisingly) a lot of misunderstanding too.

Link to comment
Share on other sites

Link to post
Share on other sites

11 hours ago, sounds said:

That's probably a bonus for now, once this project becomes more well-known it might attract the people who want something that "just works."

It's self hosted and will likely stay unsupported, there is unlikely to ever be a it just works scenario. It will be a victim to the same problem Linux has, adoption.

Google is adopted and supported because Android, Alexa because Amazon, Siri because Apple. You likely got into one of them because you already use their products and there is a good chance that other products your buying already work with them.

Willow and MyCroft have nothing to offer other than open-source and DIY, something that's is only appealing to a small number of people.

 

This will be just like MyCroft, surrounded by hype and development at launch and slowly diminish taking community support with it.

 

11 hours ago, sounds said:

 ESP-32 box is already on store shelves.

  1. I have never seen this product on a store shelf and a quick search also tells me its not available anywhere near me.
  2. It's not something that just ships with Willow ready to go, so it doesn't matter.

 

Id like to point it that I'm not necessarily saying the project is pointless, just that it's unlikely to ever be a replacement for current offerings.

Link to comment
Share on other sites

Link to post
Share on other sites

10 hours ago, Kristian Kielhofner said:

Hi, project creator here.

 

This is the thinking exactly. Willow is so new, and moving so fast, if we provided an extremely easy way to flash/install/use now it would be disastrous. The users would end up frustrated. They would end up thinking they wasted their money. They'd call it all a scam or who knows what. Perhaps worst of all, as the saying goes "you only get one chance to make a first impression".

We are catering toward users (pretty much developers, really) who look at our build instructions and think "Oh that's not bad". The build process is almost deliberately intimidating as a kind of filter to scare away those who aren't prepared or able to build, flash, and contribute. I don't want people going out and spending money on hardware that they won't be happy with because we're not quite ready for them.

 

This post from HN kind of "broke out" by making it here, to Ars, etc. There's a lot of excitement going around but unfortunately (though not surprisingly) a lot of misunderstanding too.

So awesome to see the creator on this forum. Very excited to see this project mature. I have been looking for a Alex replacement to go with my home assistant system I'm putting in my new house.

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, Nayr438 said:

It's self hosted and will likely stay unsupported, there is unlikely to ever be a it just works scenario. It will be a victim to the same problem Linux has, adoption.

Google is adopted and supported because Android, Alexa because Amazon, Siri because Apple. You likely got into one of them because you already use their products and there is a good chance that other products your buying already work with them.

Willow and MyCroft have nothing to offer other than open-source and DIY, something that's is only appealing to a small number of people.

 

This will be just like MyCroft, surrounded by hype and development at launch and slowly diminish taking community support with it.

 

  1. I have never seen this product on a store shelf and a quick search also tells me its not available anywhere near me.
  2. It's not something that just ships with Willow ready to go, so it doesn't matter.

 

Id like to point it that I'm not necessarily saying the project is pointless, just that it's unlikely to ever be a replacement for current offerings.

I completely understand what you are saying and your skepticism. I have been using Linux on the desktop since 1997 and needless to say the "year of the Linux desktop" has never happened and likely never will.

 

Mycroft is interesting but I think there are a couple of crucial differences between Mycroft and Willow:

- The Mycroft hardware is $400. That's completely ridiculous and totally out of line with the rest of the market in terms of price point. I don't know why anyone would ever buy one. The Willow hardware is actually cheaper than Echo when factoring in devices with a display, and this is before they're actually mass manufactured in earnest which should drive the price point down further (more on that later).
 

- Mycroft and previous open/open source related efforts have been failures because they made the same fundamental mistake I've seen in engineering products over the course of my > 20 year career: re-inventing the wheel and doing too much "hard stuff" at once. They did their own wake word. Speech recognition. Hardware manufacturing. These three blunders alone are the trifecta for a new effort that is doomed to failure.

Willow is a month old, and the "announcement" yesterday was really intended to be a release "by hackers for hackers". We are not even trying to target end-users, we are currently targeting developers and early adopters that will actively work with us to improve Willow. In the coming months we will have much easier onboarding and potentially even devices that ship with Willow but this is another key mistake people make - they don't understand that Wi-Fi devices are regulated and need to be certified by governments around the world. Espressif already has the distribution and worldwide regulatory stuff worked out so we will work with them and their existing distribution to make pre-flashed, ready to go Willow devices available anywhere in the world.

The ESP BOX isn't currently available because the excitement around Willow caused them to (practically) sell out worldwide in roughly 24 hours :).

The ESP BOX is manufactured by Espressif, which has tremendous manufacturing capacity. Stock has never been high because until Willow there really wasn't much practical use for this hardware. We anticipate Espressif will significantly ramp up manufacturing to meet demand.

Link to comment
Share on other sites

Link to post
Share on other sites

This is the kind of project I have been waiting for. This is the only way I'd let this sort of product into my home, and as it's running on an Esp32 that's very very familiar hardware for people doing DIY electronics and electrical engineers

I spent $2500 on building my PC and all i do with it is play no games atm & watch anime at 1080p(finally) watch YT and write essays...  nothing, it just sits there collecting dust...

Builds:

The Toaster Project! Northern Bee!

 

The original LAN PC build log! (Old, dead and replaced by The Toaster Project & 5.0)

Spoiler

"Here is some advice that might have gotten lost somewhere along the way in your life. 

 

#1. Treat others as you would like to be treated.

#2. It's best to keep your mouth shut; and appear to be stupid, rather than open it and remove all doubt.

#3. There is nothing "wrong" with being wrong. Learning from a mistake can be more valuable than not making one in the first place.

 

Follow these simple rules in life, and I promise you, things magically get easier. " - MageTank 31-10-2016

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Bananasplit_00 said:

This is the kind of project I have been waiting for. This is the only way I'd let this sort of product into my home, and as it's running on an Esp32 that's very very familiar hardware for people doing DIY electronics and electrical engineers

That's what we're really excited for too! I've been using the ESP series for as long as I can remember (back to the early 8266s) and it's a great maker platform.

 

I have even better news for you:

 

1) The ESP BOX uses the ESP32 S3, which is the highest performance of the ESP32 series.

2) The ESP BOX exposes 16 GPIOs on the back! We don't currently allow users to configure/use them but there's no reason we can't. We just want to make sure we come up with a good configuration, etc mechanism to use them.

Link to comment
Share on other sites

Link to post
Share on other sites

Can this be used to toggle wake on LAN devices?

Specs: Motherboard: Asus X470-PLUS TUF gaming (Yes I know it's poor but I wasn't informed) RAM: Corsair VENGEANCE® LPX DDR4 3200Mhz CL16-18-18-36 2x8GB

            CPU: Ryzen 9 5900X          Case: Antec P8     PSU: Corsair RM850x                        Cooler: Antec K240 with two Noctura Industrial PPC 3000 PWM

            Drives: Samsung 970 EVO plus 250GB, Micron 1100 2TB, Seagate ST4000DM000/1F2168 GPU: EVGA RTX 2080 ti Black edition

Link to comment
Share on other sites

Link to post
Share on other sites

9 hours ago, Kristian Kielhofner said:

- The Mycroft hardware is $400. That's completely ridiculous and totally out of line with the rest of the market in terms of price point. I don't know why anyone would ever buy one. The Willow hardware is actually cheaper than Echo when factoring in devices with a display, and this is before they're actually mass manufactured in earnest which should drive the price point down further (more on that later).

That just happens to be the device they sell and officially support, it doesn't prevent you from installing it on say a pi4.

 

9 hours ago, Kristian Kielhofner said:

- Mycroft and previous open/open source related efforts have been failures because they made the same fundamental mistake I've seen in engineering products over the course of my > 20 year career: re-inventing the wheel and doing too much "hard stuff" at once. They did their own wake word. Speech recognition. Hardware manufacturing. These three blunders alone are the trifecta for a new effort that is doomed to failure.

imo it really doesn't matter, in the end it does what they set out for it to do.

 

9 hours ago, Kristian Kielhofner said:

The ESP BOX isn't currently available because the excitement around Willow caused them to (practically) sell out worldwide in roughly 24 hours :).


The ESP BOX is manufactured by Espressif, which has tremendous manufacturing capacity. Stock has never been high because until Willow there really wasn't much practical use for this hardware. We anticipate Espressif will significantly ramp up manufacturing to meet demand.

No one anywhere near me that I would normally shop at offers this product for sale. I can go pick up a pi at a local BestBuy, I can't just go grab a ESP BOX.  I had never even heard of it until today. My point here was that it's not something you can just go casually buy at a local retailer. I'm not likely to buy a random item online, but I might buy it if it looks interesting and I'm standing right next to it.

 

9 hours ago, Kristian Kielhofner said:

Willow is a month old, and the "announcement" yesterday was really intended to be a release "by hackers for hackers". We are not even trying to target end-users, we are currently targeting developers and early adopters that will actively work with us to improve Willow.

I think this is the part that's being missed here. I never disagreed with this. I disagreed with it being a Alexa Replacement for everyday consumers.

 

Overall my complaint was never with the project, just how it was marketed in this thread.

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, Arika S said:

I can think of no resson why I would want a voice assistant. Open source or not, privacy focused or not.

I mean there is clearly a market for it. I would say that it seems like something that either you have a use for it or you don't so it is understandable if someone doesn't see the point in a voice assistant. It's sorta like the number pad on a keyboard. It's nice to have and some people will use it but if you don't type in alot of numbers then it might not have much use for you. 

Link to comment
Share on other sites

Link to post
Share on other sites

28 minutes ago, Nayr438 said:

That just happens to be the device they sell and officially support, it doesn't prevent you from installing it on say a pi4.

 

imo it really doesn't matter, in the end it does what they set out for it to do.

 

No one anywhere near me that I would normally shop at offers this product for sale. I can go pick up a pi at a local BestBuy, I can't just go grab a ESP BOX.  I had never even heard of it until today. My point here was that it's not something you can just go casually buy at a local retailer. I'm not likely to buy a random item online, but I might buy it if it looks interesting and I'm standing right next to it.

 

I think this is the part that's being missed here. I never disagreed with this. I disagreed with it being a Alexa Replacement for everyday consumers.

 

Overall my complaint was never with the project, just how it was marketed in this thread.

Overall, I think what's missing here is timeline and perspective.

 

The Raspberry Pi was around for at least a decade before you could go to your local Best Buy and get one. They were a pretty obscure thing for nerds for years before they got to that point.

 

We have been working on Willow for a month and it was soft announced by me to Hacker News and the Home Assistant subreddit not even 36 hours ago. If you see any other posts about it here or elsewhere, that wasn't me or anyone on my team.

 

We're a little overwhelmed by the attention it has received as our expectation was a few intrepid individuals would look at the extremely technically dense documentation, complicated build instructions, etc and maybe a few would step up and help us out. If you told me on Sunday that my HN and Reddit posts would have led to ESP Boxes essentially selling out worldwide I wouldn't have believed you.

 

I didn't have anything to do with this Ars piece or this thread. I wasn't consulted on it, I wasn't interviewed. I only found out about it by a Google Alert on my name. I didn't even have an account here until yesterday.

 

You are comparing things that are a decade old (Alexa and Raspberry Pi) to a project the world had never heard of 36 hours ago, and worse, calling for and predicting our demise.

 

We're all going to do our best to make sure that doesn't happen. Don't worry about me though - I've been in open source for 25 years. I have been told A LOT worse :).

Link to comment
Share on other sites

Link to post
Share on other sites

39 minutes ago, Kristian Kielhofner said:

You are comparing things that are a decade old (Alexa and Raspberry Pi) to a project the world had never heard of 36 hours ago.

No I'm comparing it to potential mainstream support and everyday consumer adoption. I have yet to disagree with anything else you have posted.

 

Quote

Willow, an open source Alexa replacement that respects your privacy

 

On 5/15/2023 at 7:43 PM, Nayr438 said:

If there isn't a big name behind it, it's not in store shelves, and it's not supported by the majority of products then it doesn't matter for the vast majority of people.

 

10 hours ago, Kristian Kielhofner said:

was really intended to be a release "by hackers for hackers". We are not even trying to target end-users, we are currently targeting developers and early adopters that will actively work with us to improve Willow.

 

1 hour ago, Nayr438 said:

I never disagreed with this. I disagreed with it being a Alexa Replacement for everyday consumers.

 

Overall my complaint was never with the project, just how it was marketed in this thread.

Link to comment
Share on other sites

Link to post
Share on other sites

7 hours ago, Caroline said:

Privacywashing.
What exactly makes *tihs* privacy-friendly? it says it's locally controlled but where does it get the voice recognition data from? there's got to be some service pinging the internet unless you install a 3TB database with all the samples.

You insult them and yet you don't have a clue how this works...

No, it doesn't need to ping an online service and no it doesn't need a huge DB (or a model) for wake words and a hundred or so commands in English to work offline.

 

-----

 

@Kristian Kielhofner haven't had time to checkout your project in depth, hope you don't mind me asking:
is it interfacing with ESP-Skainet or is it using ESP-SR (or something else entirely)?

VGhlIHF1aWV0ZXIgeW91IGJlY29tZSwgdGhlIG1vcmUgeW91IGFyZSBhYmxlIHRvIGhlYXIu

^ not a crypto wallet

Link to comment
Share on other sites

Link to post
Share on other sites

This sounds like an interesting project. I've never been a fan of the idea of having a device that always listens and sends data to Google/Amazon in my house, so having it running locally seems like a good idea.

The project is clearly not for me yet since it's still very early, but I am looking forward to seeing how it develops over the coming years.

 

 

8 hours ago, Arika S said:

I can think of no resson why I would want a voice assistant. Open source or not, privacy focused or not.

I think it highly depends on how you live. I don't have anything like this myself (yet) but some situations I have noticed where I would want it are:

1) When cooking and my hands are occupied or sticky, I want to be able to do things like set timers or turn on some music.

2) Quickly control lights. For example if I sit in the couch and want "movie lighting", I could just say it out loud instead of having to whip out my phone, going into an app, submenus etc.

3) Say questions out loud instead of taking the phone out and googling things I want answered when I am not on my phone or computer.

 

 

8 hours ago, Caroline said:

Yep what's the point of it?

 

And I seriously doubt it's privacy friendly, but that's what sells nowadays, apparently. Don't you think it's weird everything is private and privacy-friendly? like... download firefox, the private browser; get signal, the private messenger; use this charger, it's got a private cable. Privacywashing.

What exactly makes *tihs* privacy-friendly? it says it's locally controlled but where does it get the voice recognition data from? there's got to be some service pinging the internet unless you install a 3TB database with all the samples.

 

I get it's aimed at consoomers & oosers and *some* IT people, I'd say the sweet summer childs of IT, because the other IT segment including me don't want to even be in the same room with a smart internet of shit device, see, when you know how those things work and what corporations don't tell you about them is when you want to stay away from them. Stallman was Right, at least about that one.

Why do you not think it is privacy friendly?

You don't need a 3TB database to do voice recognition. A quick google search for "offline speech to text" will prove this, with plenty of examples.

 

I don't really see why you always act so smug and superior when it's very obvious that you often don't understand the first thing about what you're talking about.

If you don't think it works the way it says then prove it. Look through the source code or even do an experiment yourself. Don't just say "it probably doesn't work, look at how smart I am for seeing through this farce that probably fools lesser humans!".

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, Caroline said:

If you unironically believe critisicm and questioning is "insulting" I suggest you go back to elementary school or look up the definitions.

You might want to take your own advice and look up what criticism and questioning mean...

 

23 hours ago, Caroline said:

... because the other IT segment including me don't want to even be in the same room with a smart internet...

In what kind of IT are you? I doubt anything you do is IT related, I'd be surprised if you knew the difference between a toaster and a laptop.
See, I'm not insulting you, just baseless criticism and questioning.

 

4 hours ago, Caroline said:

Repfarming agrees with your friends won't make you right or anything by the way ¯\_(ツ)_/¯

When I make a mistake I acknowledge it (by replying with "my bad" or something similar), I don't deflect the subject at hand with pointless ad-hominem attacks.

Are we done?

VGhlIHF1aWV0ZXIgeW91IGJlY29tZSwgdGhlIG1vcmUgeW91IGFyZSBhYmxlIHRvIGhlYXIu

^ not a crypto wallet

Link to comment
Share on other sites

Link to post
Share on other sites

On 5/16/2023 at 1:24 AM, sounds said:

Summary

  • The good: voice detection is competitive with voice assistants from Nest and Amazon

I would only consider that as good if it doesn't trigger on anything other then its wake word like amazon and nest devices are doing currently.

Link to comment
Share on other sites

Link to post
Share on other sites

  • 1 month later...
On 5/16/2023 at 11:17 AM, Kristian Kielhofner said:

....until Willow there really wasn't much practical use for this hardware. We anticipate Espressif will significantly ramp up manufacturing to meet demand.

You're right. Using an ESP Box is a genius move here. I used it to bring my project - 'live' MLB games to life. I found it only after googling and searching to avoid all the hassles of 3D printing, PCBing, etc.  I installed and shipped to my father. Done!  Cost about the same as the esp32 DEV kit+ OLED + shipping /3D printing!

 

Willow will make support of the ESPBoxes much better!

 

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×