rhasspy VS Node RED

I've been working on this for a few months now and I think I have it to a point where I am ready to share. This is definitely a very niche solution but I am creating a new skill handler for Grocy for the Open Source Voice Assistant Rhasspy (https://github.com/rhasspy/rhasspy). My handler is here: https://github.com/MCHellspawn/hermes-app-grocy. It is not complete yet but getting there. With is skill and a working Rhasspy 2.5 setup you can do a lot of tasks in Grocy with your voice. So far you can create and delete shopping lists, create products and add and remove them from shopping lists, list chores, mark them complete or skipped, and more.

It'll run on a cheap Ubuntu box if you can't get a Pi.

And lots of people seem to like Rhasspy too:

  https://github.com/rhasspy/rhasspy

Here is one example https://community.rhasspy.org/

I tried Amazon's Alexa, the top end model with a display. Often it would taunt you about new/interesting things on the screen, but I could never get them to work. I'd had to memorize things to get even the basics working. Ended up unplugging it.

However Google's Assistant in comparison worked great, no memorization, and very useful. Sure time, weather, set timers, and alarms worked great with a very flexible set of natural language queries. Even more complex things like what will be the temperature tomorrow at 10pm, simple calculations and unit conversions. But also things like IMDB like queries about directors, actors, which movies someone was in, etc generally worked well. It seemed to really understand things, not just "A web search returned ...". Even more complex things like the wheelbase of a 2004 WRX would return an answer, not a search result.

With all that said I'm looking for a non-cloud/on site solution, even if it requires more work, most recently noticed https://github.com/rhasspy/rhasspy

I just found this one: https://github.com/rhasspy/rhasspy

Good question.

I expect that we're moving into a phase of AIs talking to AIs, and initially it'll be wasteful (because it'll be mostly English), but eventually, they'll derive their own language and seamlessly upgrade protocols when they determine they're talking to an AI. No clue how that will come about or what that language will look like, but honestly, it's kind of exciting.

Really interesting to think about how they might handle context, as well. Even though we have much bigger context windows (and they'll only get larger), context management is still a resource-management issue, which we'll probably continue to refine, as well. Imagine different strategies for managing both what is brought into the context of each request, as well as what form it could take (level of detail, additional references or commentary on it, etc). Things could get really unreadable even in English, and still be very interpretable for an LLM.

W.r.t. the graph-oriented interfaces, are you thinking something like Node-RED [1]? I'm seeing more and more people mention having LLMs produce non-text or structured outputs, like JSON, UI, and other things. Easy to imagine an LLM that wires together various open-source platforms, on-demand. Something like Node-RED for pipelines/functions, some UI tools for visualization/interactivity, other platforms for messaging, etc...

[1] https://nodered.org/

I skipped to chapter 9 in the article ("Clogged"), and it looked like Pipes failed because it didn't have a large enough team or a well-defined mission. As a result they couldn't offer a super robust product that would lure in enterprise users. "You could not purchase some number of guaranteed-to-work Pipes calls per month" is the quote from the article.

The reason I think that interesting is because that's the model these days for everything from AI tokens to Monday.com seats. It makes me feel like Pipes was before its time.

That said I've been collecting different "business glue" products that are similar to Pipes. To me, like you say, they aren't as interesting, exciting and intuitive as Pipes was, but maybe it just takes a little more digging. I tried to focus on open source tools but some aren't.

- n8n io: https://n8n.io/integrations/mondaycom/

- Node-RED: https://nodered.org/ (just read about this one in this thread)

- trigger dev: trigger.dev

- automatisch.io: https://automatisch.io/docs/

- Activepieces: https://www.activepieces.com/docs/getting-started/introducti...

- Huginn: https://github.com/huginn/huginn

- budibase: https://budibase.com/

- windmill: https://www.windmill.dev/

- tooljet: https://www.tooljet.com/workflows

- Bracket: https://www.usebracket.com/pricing (just SalesForce <-> PostgreSQL)

- Zapier: zapier.com/

Anyway I hope some of these are fun!

Source: https://github.com/node-red/node-red

I believe Node-RED (https://nodered.org/) the way to go. It's just an NPM package to install and you can run it how ever you wish (even on Windows). It has a friendly and helpful community with even the main developers tirelessly answering even beginner level questions. In fact the community forum its THE friendliest forum I've ever been a member of by a large margin. Node-RED's development is supported by the JS Foundation and it's completely free and open source. It's widely used in the industrial automation industry and even integrated by some PLC manufacturers such as Siemens.

Does anyone know if their are plans to implement something like this (or if there are already custom nodes out there). I'd like to experiment with things like looping and incrementing values (like a for loop) for a Ksampler for example. It's only an example though, so I am not looking for a ksampler specific solution; just a generic way to have a variable (e.g. Seed value), run some nodes that use that value, increment the value, and then loop back to the beginning until some sort of condition is met. Node-Red (an event driven node based programming language) has this functionality so it could defintely work in a node based environment such as ComfyUI (see here).