157 comments found.
The product is great! But the best thing is the service, always ready to respond quickly and with a good disposition! Absolutely recommendable. Thank you!
Thank you for the review and the kind words! It is a pleasure to work with you too!
Regards
Hi Nick,
How to disable asking for name at the start? I want customer to just start Chat by greeting (without name) from an avatar.
Thank you
Hi Andy,
For now this is not possible to remove, as browser is requiring the user to interact. I am checking for a possible solution.
Regards
Ok. Thank you
Hi Nicky75,
This is really an amazing and interesting product.
But with the couple of options and addons available I’m confuse about it. Can you please guide me, if I want the achieve similar to this demo for a small business website?
I’ve a WordPress website, I want to integrate an avatar which talks with customer and reply from pre-defined set of answers related to the business. The site is currently hosted on a shared server with good resources with ssh, node.js and python.
I appreciate if you can explain which of your application or bundle is required and if any other additional requirement on server, ChatGPT a/c and any other tools or plugin needed to make it work.
Thank you
Hello,
Thank you for your kind words!
You need just this product – https://codecanyon.net/item/livesmart-ai-video-smart-video-avatars-with-chatgpt/52246761/ and no other bundles or plugins. It can be installed on shared hostings under a subfolder for example with only one requirement – PHP version at least 8.1. Of course you need ChatGPT (OpenAI) and HeyGen API tokens. Check please this article and the video at the bottom on how to obtain them.
You can integrate LiveSmart into your WP, there is a WP plugin inside the extension folder. Here is a demo on how this looks like on WP page.
Email me if you have more questions or doubts.
Regards
Thank you
Can the generated avatar be downloaded?
Hi Nicy, What an incredible product, congratulations. I want to make an AI Psychologist and an AI Astrologer, I have a ready-made SaaS, but I’m thinking about switching to yours. Some questions: 1 – ChatGPT does not retain basic information and does not retain the last interactions, which is essential for the work. This problem can be solved with some APIs on the market or by working on the script. I want to know what you have in the script that can solve this problem. Thank you very much
Hello,
Thank you for your interest in LiveSmart and kind words!
1. ChatGPT keeps context of the conversation but in the current session. Do you want it to be kept per user for days ahead? It may reduce performance.
2. There is a history of the conversation per user (only identified by name). It can be used for keeping the context not only for the current day session.
Login to https://www.new-dev.com/videoai/dash with demo/demo, create a new room from Rooms – Video AI Avatars and in Avatar attitude field add your instructions. Or send me a PDF file, so I can upload for ChatGPT assistant and you can test it. You see that in your current conversation the context of the chat is kept.
Regards
625 / 5.000 Hi Nick,
Thanks for your answers.
2. There is a history of the conversation per user (only identified by name). It can be used for keeping the context not only for the current day session. Me: ChatGPT stores the history of the context of conversations during a session, that is correct. But if you log in again, all this history disappears. I have already done several tests with my API.
It is still not clear to me if your script retains: basic information and the last interactions with the user, to continue an interactive and concise conversation after a new login.
Please let me know your email.
Thank you very much
Hello,
No, the application is not keeping context out of the current session. But I will add this as a feature request.
Regards
Hi, I am trying to understand how many cost involve here. OpenAI, what is the role of OpenAI? is it for the voice or for input and output interpretation ? Heygen for interactive avatar, mean lip movement? The heygen costing is a bit confusing, there are free avatar and heygen charge $0.10 per minute, so mean whenever the ai avatar speaks, we are paying $0.10 per minute plus openai credit?
Can I replace OpenAI with open source LLM such as Mistral or LLama ? If I can , it will reduce dramastically the cost since hosting own LLM is cheaper. If open AI is for the voice, is there open source voice LLM to replace openai voice?
Is there any heygen open source replacement? 1000 minutes for $100, i am afraid, 1000 minutes could only last a few days and eventually will cost us thousand $ for this heygen.
Hello,
Thank you for your interest in LiveSmart!
Here is how the product works – using voice recogtnition, your query is sent to OpenAI (using chat completion, tools calling or OpenAI assistant, this can be configured in the Dashboard). The text response from OpenAI is sent to HeyGen avatar and it do the voice and appearance. It is not only lip syncing and you can change the voice, language and attitude runtime or from the Dashboard too.
For now I am using only OpenAI and HeyGen combined. In near future will add some alternatives for both.
About pricing – OpenAI prices are low, the demo is based on gpt-4o and using assistant also. For this month till today charhe is $3.62, not much. HeyGen is charging only when the avatar speaks, not the entire session. I have multiple demos – about 20 per day in fact and still not spent all the 1000 credits.
For me this is acceptable. Please email me if you have other questions or doubts.
Regards
Thank you for the details explaination.
Can you share what openai model are you using? i heard 4o is among the cheapest.
While for the heygen, i tested your demo and there is no avatar, only sound, so i assume if you were to activate the avatar, the cost will be double of your current charge? By the way, for heygen, is it charge per second or per minutes for the session? what if the session is 45 second or 1 minute and 45 seconds. how will be credit be charged.
by the way, for the language selection, can you also add option for user (not dashboard) to select. maybe a flag icon beside the head or maybe like a pin on the shirt with an arrow icon where when click it will popup a selection of languages.
I found this https://github.com/ali-vilab/dreamtalk it seems interesting and maybe open source, mean we could implement it and could replace heygen for the long term.
Hello,
I am using gpt-4o model.
You should see the avatar. It does not depend on the token. My server is a bit slow and I changed the time for avatar appearance. Please check again.
The charge is per seconds, not rounding to highest minute. If you check my screenshot, there is a decimal in the remaining credits.
There is an icon for language change.
There are plenty of video generating AI tools, but no one is doing a streaming like HeyGen for now.
Regards
Nick really i never see a so perfect Support, thank you so much for your Help. The Script running now Perfect 
Thank you Nick great Work
Thank you!
Hello, how many times can the purchase code be used? What I mean is on the same server and domain
Hello,
The purchase code is assigned to one domain. You can have one purchase code for one instance. But if the case is that you need a test and a production environment, you can email me from the support form where you need it enabled.
Regards
Hi when will be the next update please? You mentioned features on version 1.0.7 that interest me. I wanted to wait for the update before buying
Hi,
Most probably tomorrow or this weekend. I am doing some more tests.
Regards
Hi, ok no problem thanks.
Hello,
I have just released 1.0.7 version of the product.
Regards
Hello, ok received, Thank you.
Hi, if I buy the script can you help me with the weather and news? But I want the news to come from my website and I don’t use rss feeds.
Sure will help you. News are coming from an API. Newsfeed demo is using RSS feed. Do you have an API endpoints for news and weather?
Thanks nicky, no, i don’t have an API endpoints for news and weather. That’s the problem
Well, if your site has news and weather information in the DB, you can fetch directly the information and send to the avatar to speak.
Check please the getAvailableTimeslots and bookTimeslot from the API documentation.
Yes, the news and weather information is in the database.
Then you can directly connect to the DB and retrieve news and weather from the DB. If Video AI is on the same server, it can connect to the news feed. Booking demo is doing just this.
ok no problem so for the news we will use the database of my website and for the weather we can even use the api that you used in the demo But I don’t know if you noticed on the video it doesn’t predict the future like tomorrow or a later date it’s only today. Can you manage that for me please so that it predicts the weather in the future like tomorrow, the day after tomorrow or a specific date? Thanks
OK, I modified it, and you can include date also. For example “Weather condition in San Francisco for Friday”. If you want current weather condition, you can ask “Current weather condition in San Francisco” 
whaou, That’s great, thanks Nicky. We’re going to do that then for the weather we use the API that you used for in the demo and for the news the database of my website
Ok, ping me if you need more assistance.
No problem, thanks Nicky
Hi, my server and database were updated yesterday and the outgoing communications are changed. The site won’t work properly as is so I need a list of your outgoing connections/api url’s for the site so I can whitelist them or the site won’t work properly again. Thanks
It was just the heygen api I needed to whitelist and it’s sorted now
Cool!
Hi, great work and when everyone is only talking about AI in content generation, this product seems a bit different and promising. I have few queries related to the current product.
1. Is it SaaS enabled? I see you have mentioned this in your July 15 update but I couldn’t find it. Also, is that included with the regular license?
2. What are you planning to include in the future updates?
3. Are you planning to include different integrations and automations like with CRM, Conference tools, Video streaming platforms (YouTube) etc?
4. I found it very slow to respond and if a speaker gives a pause in between the sentence it doesn’t understand. Have to say it very clearly for it to understand. It might struggle where people doesn’t have fluency or have different accents.
5. Do you have any plan of adding a landing page to be included with the SaaS package?
6. Can these avatar be a bit more lively because right now they have a very flat tone, expression, and body language.
7. With different languages can we localize the Avatars as well? like in terms of look and feel of the avatar?
Hello,
Thank you for your kind words and your interest in LiveSmart!
1. Yes, it is SaaS enabled. You can read more about monetization in this article, how to test and what the admin UI looks like. You need Extended license to have SaaS enabled, but you can first test it on Regular and if you decide to go with Extended, I will do a discount.
2. In the coming 1.0.7 release is planned:
- Added formatting of OpenAI response;
- Profanity filter;
- Payment per organizer, not global;
- Feature request – added interrupt button;
- Feature request – added run-time change of avatar voice;
3. LiveSmart AI Video can be integrated with ease anywhere. Here is a demo how is included in a WP site. Also there are a lot of sample integrations with booking, news, weather, currency conversion platforms. If you need specific integration with a specific platform, please email me.
4. Yes, you must speak distinctly and clearly in order speech recognition to work as expected. Also please have in mind if you are speaking in other than English language, to choose the language from the top-right drop down menu while in a meeting, or to create own room with native language from the dashboard with demo/demo.
5. Sorry, what landing page? You mean to show subscription plans and so on? They are included in the dashboard and created by the admin. Check the article for the monetization, there is a video there.
6. Yes, you can choose from the menu top-right, what to be the attitude and language of the avatar to be – there are different types of English for example – cheerful, calm, professional, etc.
7. Yes, login to the dashboard with demo/demo, go to Rooms -> Video AI Avatars and you can choose from different avatar types. There are 15 free avatars. For the paid ones you need a subscription plan from HeyGen – the provider of the avatars.
Email me if you have more questions or other concerns!
Regards
Hi I hope you are well? you told me recently “Files can be used with the assistant for now. But you can train the avatar with the text field and / or with adding custom functions, like booking, weather and news demos.” is it possible for the news for example to connect the avatar with a news website for example so that he or she reads us the news like in the demo or how does it work exactly with the news? Thanks.
Hello,
Now you can assign an assistant from OpenAI to an avatar. If you create a new room from the Dashboard with demo/demo. You can email me a sample PDF file from the contact form, I will create an assistant and assign the file to it. When you login to the dashboard, you should see the assistant in the dropdown.
About news functionality, it is not using an assistant, but the advanced functions. There is a sample function for it, you can see the API documentation – getNews and getNewsDetailed. Here is below how the news room is set:
Ping me if you need more clarifications or have other doubts.
Regards
Thank you nicky for the explanation it is clear. but one thing please. the news comes from “https://eventregistry.org” or I can put any source for example CNN or if I have my own news site?
You can put any source there, but as this is a sample function and it is parsing the response from eventregistry, you need to take care of the response too. Have in mind that these are sample functions on how you can extend your avatar interaction. They can be used in real life of course.
Regards
ok i understand now and thank you very much.
Hi, are we able to use GPT-4o-mini ? and do you have a roadmap for updates ? Thanks
Hello,
Yes, you can add any model, including gpt-4o-mini. The coming release 1.0.7 will include:
- Added formatting of OpenAI response;
- Profanity filter;
- Payment per organizer, not global;
- Feature request – added interrupt button;
- Feature request – added run-time change of avatar voice;
All these can be seen in the demo pages.
Regards
Can we change the characters ourselves? And they only support English
hi great script again. 1) I tried with the OpenAI Assistant in the demo and as you said it takes a while to respond.2) And I find that by training avatars and limiting it a little you can, for example, increase knowledge with files, website links and others. 3) Also if you ask a long question, it takes a long time to answer even if the answer is already in the chat and in addition the answer is displayed and it still takes a lot of time for that speak. Thanks
Hello,
Thank you for your kind words!
1. Yes, Assistant API is a bit slow. In order to speed it up, you can use tier 3.5, not 4. Use simple PDF or TXT files for better reading. Use 10-20kb files. Also you can not use the assistant, but use the instructions field from the Dashboard and paste plain text there.
2. Files can be used with the assistant for now. But you can train the avatar with the text field and / or with adding custom functions, like booking, weather and news demos.
3. I have added the text sent to the avatar to be splitted into chunks of 800-1000 characters. But you can set this to 200-300 characters and the avatar response will speed up.
Email me if you have more questions or other concerns!
Regards
Hi thanks hope you are fine, OK thank you
Hi, I am interested in purchasing the script. However, I want to check if there any subscription fee to heygen in order to have the video Avatar works. I noted that heygen charge 1 credit for 1 minutes video, say user use the video chat avatar, then I must pay for each minutes of usage.. Is this correct? Thanks, Dan
Hello,
No, you can go with free HeyGen token without a problem. My demo site is with it.
LiveSmart model perfectly fits for free HeyGen token – interactions with avatars, 15 free avatars, three sessions at once, 1000 characters of speech at once and over 140 localized voices. Check this at our demo site with demo/demo credentials.
HeyGen charges 1 credit for 1 minute of uninteruptive avatar speech. If you need more avatars or voices other than the listed over 140, then you can go with a subscription plan. If you need a custom avatar, HeyGen is offering it for $500 one time payment.
Email me if you need more clarifications or have other concerns.
Regards
Hi excellent script. Question before purchase: I noticed that if you ask a long question, it takes a long time to answer even if the answer is already in the chat and in addition the answer is displayed and it still takes a lot of time for that speak.
Hello,
Thank you!
Yes, when there is a big text to be voiced there is some delay about 3-4 sec before the avatar starts speaking. That is why you need to give instructions to the avatar to answer with short sentences. In the coming release I have added a possibility to chunk big text into small pieces to ease the avatar speech.
Email me if you have more questions or have more concerns.
Regards
Ok thank you very much one more thing please. Does the memory feature exist? or we have to start the conversation from the beginning every time? THANKS.
Yes, by default the avatar is keeping context of the conversation. For example you can ask it “Where is the oldest tree in the world?” and after the answer – “How can I reach there?”
ok I see but when we leave the conversation is the memory erased?
Context is kept per session not per room. When you exit and re-enter conversation starts from begining. On a later stage may add some login information to keep the context per user.
ok I understand, it’s cool thank you.
Hi I sent you an Email.
Hello,
I have just emailed back
Regards
Does it have a platform translation module? You already have Brazilian Portuguese
Yes, you can localize the video panel and the dashboard.
In the demo version there is no option to view the panel in Brazilian Portuguese, is it possible to enable it?
Hello,
Check please this avatar if it is native. I will add locales in the voices, as they are all listed Portuguese, but the locale is missing. This is pt-BR.
Regards
I have updated the dashboard, you can check. The change will be included in the coming release.
how i can i change build js file? where i will find unbuild js file?
Hello,
Sorry, didn’t get this. You need to check unminified version of the JS file? Actually the communication with OpenAI and HeyGen is handled by the PHP back-end script files and not from the JS.
Regards
Hi, just wondering, was that an update on the 8th or not ? as nothing showed in admin. Just want to make sure.
Hi,
It was an emergency fix and it was only for new users as SQL sequence was breaking the install process.
BTW SaaS version is uploaded to demo site, together with some fixes and UI rearrangement. You cam login with demo/demo and check. Here is how it looks like for admin
Regards
Thanks, SaaS looking good too.
Thank you!
Hello,
I have just updated to 1.0.6 release with SaaS and monetization module included.
Regards
