The r1 needs webhooks. If this were to happen I’d find a way to integrate with Hubitat and have rabbit control devices in my home.
Yep. We’re hoping to launch them sometime this month (June). Don’t have an exact ETA yet but it’s looking likely to happen before the end of the month.
I imagine webhooks are trickier than most people realize when dealing with intent and the finicky nature of the current state of LLMs
I understand. Me as a tinker (and i’m sure other people) don’t care as much about a finished product. I just want something.
I hear ya, but I also imagine the immediate backlash on X / Discord of people complaining it’s not finished. Which is sad.
Honestly it’s going to be great rabbit has a limited number of developers but there is lots of people who can code so any ability to extend rabbit is only going to make it more attractive for everyone.
any updates on this? I am truly looking forward to the possibilities that webhooks will unlock.
If the use of this device is going to scale, you need to open up the tools environment to all web developers to integrate their UI. I have access to several proprietary APIs that I would like to support. When and how will I be able to integrate my apps.
I am quite happy to hear that support for webhooks is coming soon. I love the LAM idea of learning to use UIs but this would be amazing for developers in the beginning while teach mode is still being developed. And even once teach mode is completed, I still feel that the eventual goal is having API support for all applications that the R1 or any other AI agent interacts with because it will provide more deterministic behavior. I cannot wait!
Hi Guys apis or webhooks are going to be important. In based on Australia and the only function working effectively does the same thing as Seeing AI plus translator. We will need either a basic rabbithole settings function for localisation and a developer platform (usong r1) that can build integrations.
I was hoping that Webhooks would came with the latest update, this as it was a combination with OTA and web.
@simon please tease us with a new expected date
I don’t have an ETA just yet but it’s still very much in development
Any update one this?
I think this is the single most useful feature that could possibly be implemented. It’s would also be the highest yield utility gain for the amount of effort.
Sure getting it to interact properly with other tools adds complexity, but even just the ability to send our user messages as is to a webhook would unlock an insane amount of use cases even for the no code guys. Webhook in make or zapier with an llm step to infer and direct to any tool / automation. I’m sure there are more user friendly options that require more development like keyword or phrase to direct only certain messages to webhook/certain hooks.
Loving the playground though, you guys are killing it!
They have said recently that it has been shelved for now to get playground and teach mode running. They also want to release both to determine if there is still desire to have webhooks on top of all the functionality they have planned for the other two.
From what I can see - the feature to edit a note via the rabbit hole has gone? - so would really like to see webhooks become available!
Notes are kind of unrelated but they said you currently can’t edit new notes because they had to change the system to handle markdown so they will be adding editing back in the future.
I’d personally like to see us still do it, but I’m not aware of any active development on it at the moment - majority of resources are on improving playground and getting Teach Mode to Beta stage so that we can open it up to more people
There will still be massive interest and utility for it. A browser automation however smart isnt as powerful and fast as having a direct webhook that can follow whatever paths we set out for it to follow.
Worst case I guess I’ll teach it how to interact with postman lol