Have you ever wished to have a version of yourself in multiple places at the same time, talking to various people?
While it may not be physicall...
For further actions, you may consider blocking this person and/or reporting abuse
This is awesome 👍, but my question is that , from the first step, it seems to me that it is ChatGpt that constructed the numbers of questions to be ask which inturn at the end of answering these questions generates a transcript for the users.are my right? If yes, is there a way will can Increased the numbers of questions attempts to let's say 10 or morthan the 7.? Or 7 is the default?
Yeah its totally up to you if you want to increase the questions or personalise it, the questions mentioned are one way to do it you can customise the questions as per your need and as per the data you want to feed in
Where can that be achieved? In the same side bar where one generate the API key ? I mean Do ChatGpt?
So for that you can copy the same prompt from my chatgpt link and paste the prompt in your chatgpt with changed questions
Let me rephrase my question;
What I mean is that,if I what to customize my questions prompt, where can I do that? Same DocGpt where you obtained the API key?
Yes you can do it in DocsGPT where i got the api key
Right now, I'm inside the app.docsgpt.cloud
Where can I click to customize my questions, I'm seeing, source docs, settings, upgrade to pro etc
Hey you can reach out to me on docsgpt discord I could help you
better there : discord.gg/Bbek2pzH
I am almost done , i have a test server where i hosted my test apps, can i show you the configuartion, i have been able to embbed the widget into my website , but once i initiated a chart, it will trigger no network, i want to show you the screenshot of my configuration
Tell me what is wrong here, pls
I guess there is the problem with the api host url, check that once
i used my domain name url instead of 'localhost:7001',
Are you trying to deploy it somewhere or are you just trying to test it by your html
Yes, I am actually trying to deploy it into my website,
Whoa! It sounds so awesome to interact with my own twin haha 🔥
Haha true, i use it to interview myself
That sounds pretty interesting! It's just a bit unfortunate that it won't actually be a totally accurate clone.
If the API key is so secure, why isn't it encrypted in the script block? Couldn't anybody steal the key and find out information about you?
While the AI twin can’t fully replicate a person, it does capture enough to create meaningful interactions. As for security, the API key shown is just an example. In production, keys are typically stored securely on the environment variables.
Thank you for explaining!
It's an interesting experiment :)
Definitely! It’s amazing to see how far we can take this idea!
But this twin does not work 24/7 without you controlling him right? Would be awesome to have a agent working on tasks, reacting to events that might occur without you doing anything. Doing it in a way you would may be a benefit (or maybe not).
Agreed, but in the chatbot case it works 24/7 answering various queries and as you mentioned the possibilities to expand this idea are endless with some modifications and an idea anything could be done like doing tasks and reacting to events
Creating a virtual AI twin 🤖 is a good way to extend your presence 🪄 on platforms 24/7. This guide makes the idea accessible and actionable 💯.
It can also be used as a Telegram or Discord bot 🤖🪄.
True 🔥
Love this idea! Such interesting work! 👀
Thanks 🌟
This is an awesome use case! Can my AI twin do coding and schedule/take meetings for me as well? 😉
Haha well it depends on how well you could customise it according to your workflow but if not managing your schedules it can always be someone whom you could talk to 😉
To create a digital twin like the author did, start by collecting data on your personality and responses. Train your AI using tools like DocsGPT to simulate your communication style, then set up a personalized prompt for authentic interactions. Finally, deploy it using an API to have your virtual self available on multiple platforms!
Damn!
Cool. Thanks for sharing!
Glad you liked it 🌟
Is there a danger that it can be used for exposing some unwanted informations about yourself? Not all users on the internet are good.
Hey, yeah as per the prompt it says that the personal data shouldn’t be mentioned in the response
Nice, but I wonder how can you do the same thing but with an open-source LLM on your own server...
Setting up a digital twin with an open-source LLM on your own server is very doable. You’d start by hosting an open-source model like LLaMA or GPT-NeoX on your server, then train it on your data using a fine-tuning process. With this approach, you get full control over your data, customizations, and security. Plus, it allows you to avoid third-party dependencies.
Thank you
Welcome
Would it be possible to do this entire process through the API, so that I could automate the creation of these chatbots? Or can you only make and train the chatbots through the API?
Also, can you turn on memory for one of these chatbots? So that each chatbot remembers its previous conversations?
You can only query chatbots via an api, yeah for this you will need something like a global api key, we have not added it yet. To use memory you have to log and store it when sending request and just adding it inside the history param