What i am missing in your videos of chatbot like a live website WordPress, making a chat tool of that website API, with example 2 testing tools (order status, order return) with live RAG chatbot working, so its could be a complete flowise package ,as RAG its working fine but on tool level is complicated example: as a user i came and ask about my order, that tool brings relevant information about that order or its status
Truly amazing Leon, very useful and easy for non-coders. It will be wonderful if you make the voice assistant to work with Flowise. Not only your videos are explained in detail and easy to understand and follow, but the topics you select are just perfect - real life applications! Thank you Leon! ⭐⭐⭐⭐⭐
Looks a great product. The only minor concern I can see is with the knowledge bases. Unless they have RAG / Vector DB under the hood somehow, it looks like they are passing the entire KB in each prompt which would be expensive and inefficient for larger knowledge bases. Great video, as always.
I've been having issues getting the data into the Webhook - maybe there were other settings in that webhook node on N8N that is preventing those 4 keywords from populating my airtable. Any thoughts? Thanks! Great tut.
The quality's good, but the pricing doesn’t really make sense. At 500 credits/month for testing and 1000/month for production, it’s about $0.22 per minute + LLM costs-way higher than Retell and Vapi. And since they both use ElevenLabs too, I’m not really seeing what sets this apart.
Agreed..it doesn't make sense. way cheaper to just use their api through one of the services you mentioned or through your own app. I don't understand the advantage given the price.
Completely bonkers. They price it to compare to a human cost. Hope there will be huge competition soon to get us to 2-5c per minute. Then it will blow up.
Great overview! I saw that they have a Client Tool as well that can work with the user's DOM in the browser when it is triggered. This would be an awesome follow up video. I would like to see the cool things you can do with this! 👍
Leon, I truly enjoyed your video. I have a suggested enhancement to the reservation part of your agent. You might want to create a database the lists how many tables of each size, that you have. This way, the agent will know if all the tables are full when someone makes a reservation. Just a thought! Thank you for an informative and interesting video! 🕊️🫶
I think you could make such an assistant locally Leon, you would use TTS or XTTS using flowise or n8n. The question is how to get such fast audio to text processing. A local Jarvis who answers questions based on the local knowledge.
Yeah, it's not that simple. I think the difference is the latency, streaming and ability to interrupt the conversation. Also, this is super easy to add to your website. Would definitely be cool if Flowise could include something like this in their web embed?
very easy to do but, as mentioned in another comment, I don't know why somebody wouldn't just use their api for way cheaper then this service. If you're a no-coder you could spin this up in retell, vapi, etc for less cost.
Agents can use "tools" to interact with the worlds around it. You do need some experience with webhooks / API's to understand it, but it's well worth learning if you want your agent to do more than just simply generate text from a knowledge base.
Thanks Leon. Very well done. What if my chatbot is based on super complicated flowise chatflows or agents and i want the conversation to be about about what those chatflows can generage ?
This is basically the audio equivalent of text streaming with LLMs. We no longer have to wait (sometimes much longer than a second) to receive the final response, making it feel more natural. And, as with text streams, we can interrupt the respondence.
Bro i am using this feature just for fun. Could you provide details like about bill cycle? It is postpaid subscription? Because it say : in-development calls are billed at a rate of 500 credits per minute
Hi Leon, thanks for this great video! It seems ElenLabs is stepping into the AI Agent space-exciting! Quick question: how can we make the agent check calendar availability? Also, I like your smooth and clear explanation style; it would be awesome to see your take on Bolt.new and Windsurf! 👏🏻
Hey, thank you the kind feedback and suggestions. You can use platforms like n8n or Make to create the calendar logic (like I did for the reservations in this video).
Great video Leon, as usual. Do you know how many bots you can have at a time? For example if I am on the Creator Plan, which is 100 minutes, can I have say 3 bots but cumulatively they have 100 minutes? I cannot seem to find that on the site
@@leonvanzyl Thanks good point, but for South Africa dollar exchage rate is a killer. At some point this has to reach a commercial rate, and I doubt Elevenlabs is going to deviate much from their current pricing model. Still very impressive tech no doubt. Google has a REST API and SDK's for Python and other languages so it's fairly easy to implement.
Not quite accurate - you only pay when the agent is ‘working’ whereas a human worker is paid by the hour, irrespective of how many minutes in each hour they are productively adding value to your organisation.
Try this yourself.
try.elevenlabs.io/up77070vqvjl
Enjoy the video! Want to see more tutorials on this? Perhaps Python / TS SDK?
What i am missing in your videos of chatbot like a live website WordPress, making a chat tool of that website API, with example 2 testing tools (order status, order return) with live RAG chatbot working, so its could be a complete flowise package ,as RAG its working fine but on tool level is complicated
example: as a user i came and ask about my order, that tool brings relevant information about that order or its status
Truly amazing Leon, very useful and easy for non-coders. It will be wonderful if you make the voice assistant to work with Flowise. Not only your videos are explained in detail and easy to understand and follow, but the topics you select are just perfect - real life applications! Thank you Leon! ⭐⭐⭐⭐⭐
Thank you for the amazing feedback 🙏
Great explanation, i hope also with a intergration Flowise and Elevenlabs.
Looks a great product. The only minor concern I can see is with the knowledge bases. Unless they have RAG / Vector DB under the hood somehow, it looks like they are passing the entire KB in each prompt which would be expensive and inefficient for larger knowledge bases. Great video, as always.
I'm pretty sure they're using RAG as well. It's kinda the standard
AI that simplifies, automates, and transforms 💥
Love to see the growth! Keep going, Leon.
Hey Richard, it's awesome seeing you in the comments!
Thanks for the support 🙏
Absolutely exceptional! It sounds perfect...a little too perfect. Make it sound lazier and disinterested, and then you've got yourself some steak!
Bwahahaha!
I've been having issues getting the data into the Webhook - maybe there were other settings in that webhook node on N8N that is preventing those 4 keywords from populating my airtable. Any thoughts? Thanks! Great tut.
The quality's good, but the pricing doesn’t really make sense. At 500 credits/month for testing and 1000/month for production, it’s about $0.22 per minute + LLM costs-way higher than Retell and Vapi. And since they both use ElevenLabs too, I’m not really seeing what sets this apart.
It's free to use during the beta phase, so still work checking out.
Agreed..it doesn't make sense. way cheaper to just use their api through one of the services you mentioned or through your own app. I don't understand the advantage given the price.
Completely bonkers. They price it to compare to a human cost. Hope there will be huge competition soon to get us to 2-5c per minute. Then it will blow up.
Great overview! I saw that they have a Client Tool as well that can work with the user's DOM in the browser when it is triggered. This would be an awesome follow up video. I would like to see the cool things you can do with this! 👍
Agreed! Been looking at the React SDK.
Amazing and well detailed video! 👏👏👏
I’m working on a complex AI service for real estates and I wondered if I can get your support on it?
Great video! Could you make a tutorial on how to integrate this into a game, say made with Unreal Engine?
Brilliant video. Super useful. Thanks for sharing
You're welcome
That's amazing. Just checked it. Thank you, Leon
I love this kinda tech. It can only get better (and cheaper) from here on out.
Just great, once again thank you for the amazing content!
You're welcome 🤗
Leon, I truly enjoyed your video. I have a suggested enhancement to the reservation part of your agent.
You might want to create a database the lists how many tables of each size, that you have. This way, the agent will know if all the tables are full when someone makes a reservation. Just a thought!
Thank you for an informative and interesting video! 🕊️🫶
Absolutely!
This was just an overview of the platform itself, but you can definitely build that integration in the same way.
Leon, please make a video about ElevenLabs and Vectroshift voicebot integration.
Great idea, thanks!
@@leonvanzyl welcome
I think you could make such an assistant locally Leon, you would use TTS or XTTS using flowise or n8n. The question is how to get such fast audio to text processing. A local Jarvis who answers questions based on the local knowledge.
Yeah, it's not that simple. I think the difference is the latency, streaming and ability to interrupt the conversation.
Also, this is super easy to add to your website.
Would definitely be cool if Flowise could include something like this in their web embed?
Excellent work Leon - As an extension to this - would be nice to integrate the AGENT in to telephone calls...I believe this is possible.
It definitely has the potential!
very easy to do but, as mentioned in another comment, I don't know why somebody wouldn't just use their api for way cheaper then this service. If you're a no-coder you could spin this up in retell, vapi, etc for less cost.
Mindblowing Leon! I fall in love with you 😅
Thanks 😊
Thanks for sharing the knowledge
You're welcome 🤗
great video! Which microphone you use for making video?
Thanks!
Not every day someone asks about my setup 😁.
It's a Shure SM7B.
@@leonvanzyl Awesome man you inspire me !
What can be a good cheaper alternative for me starting out?
hey thanks for this tutorial.I
What are the costs ofr using such service (from 11labs ?)
It's free during the beta phase. They do mention pricing on the website
Well-explained 🙌🙌
Thank you
I did not understand that tool thing, can you explain from scratch
Agents can use "tools" to interact with the worlds around it.
You do need some experience with webhooks / API's to understand it, but it's well worth learning if you want your agent to do more than just simply generate text from a knowledge base.
Thanks Leon. Very well done.
What if my chatbot is based on super complicated flowise chatflows or agents and i want the conversation to be about about what those chatflows can generage ?
I would love to see this being integrated into Flowise.
Great I will test it:) thank you
Have fun
Why use this when we have whisper and deepgram. Is 1 second delay really that important??? Is this what market needs really???
This is basically the audio equivalent of text streaming with LLMs.
We no longer have to wait (sometimes much longer than a second) to receive the final response, making it feel more natural. And, as with text streams, we can interrupt the respondence.
Very impressive! How much did this conversation cost, Leon?
I think this was still within the free tier.
Bro i am using this feature just for fun. Could you provide details like about bill cycle?
It is postpaid subscription?
Because it say : in-development calls are billed at a rate of 500 credits per minute
I read somewhere (in their documentation) that it's free during the Beta phase.
I don't think they've confirmed the costs yet.
Leon question how to locale and teach the agent that is 2024? im from mexico and when I try to make a reservation the agent tells me that is 2023 jeje
Add a simple tool to the agent that retrieves the current date and time using JS or Python code (use chatGPT to generate the code for you)
Hi Leon, thanks for this great video! It seems ElenLabs is stepping into the AI Agent space-exciting! Quick question: how can we make the agent check calendar availability? Also, I like your smooth and clear explanation style; it would be awesome to see your take on Bolt.new and Windsurf! 👏🏻
Hey, thank you the kind feedback and suggestions.
You can use platforms like n8n or Make to create the calendar logic (like I did for the reservations in this video).
Really cool! I wonder what an average interaction will cost?
It's free at the moment (while they're in beta). It's possible the pricing model will change based on things like the LLM that you use.
Great video Leon, as usual. Do you know how many bots you can have at a time? For example if I am on the Creator Plan, which is 100 minutes, can I have say 3 bots but cumulatively they have 100 minutes? I cannot seem to find that on the site
Hey Wayne, thank you for the feedback.
You can have as many bots are you want. They'll collectively use your available minutes.
Amazing 🤯🤯
I wish I could show more. The game character stuff is hilarious
Hi Leon, quick question, does it not work on local host? Must I publish my page first?
Hi Wayne, I assume you're referring to the website itself? I ran a simple HTML page locally in the demo.
@@leonvanzyl I think its the first two levels where you cannot publish your bot, only from creator level upwards.
Amazing 😮
I know, right?
@@leonvanzyl yeah, and your content is amazing too
wow , this so cool
It's really impressive
Can we combine this with our flowise chatbot?
Oooh, now that would be cool!
Yes, wonder the same thing, especially integrating into a multi-agent system would be awesome. Thank your@leonvanzyl.
Is it possible to ban/block an IP address the excite a certain length of conversation?
You should definitely check the security section
SWEET
COOL
🤣🤣🤣 Leon you kill me
Something I said? 😊
This is an awesome adaptation
I think it's better than vapi!!
It's definitely easier to implement.
This is cool but expensive. Gemini 2 can now do the same, probably be better priced.
It's free during the beta phase. Worth checking out in my opinion.
Not sure that Gemini 2 makes integration as easy.
@@leonvanzyl Thanks good point, but for South Africa dollar exchage rate is a killer. At some point this has to reach a commercial rate, and I doubt Elevenlabs is going to deviate much from their current pricing model. Still very impressive tech no doubt. Google has a REST API and SDK's for Python and other languages so it's fairly easy to implement.
Cost about the same as a real person per hour
Not many people are willing to work 24/7 without leave 🤣
Not quite accurate - you only pay when the agent is ‘working’ whereas a human worker is paid by the hour, irrespective of how many minutes in each hour they are productively adding value to your organisation.