I have recently found a very good tutorial by @nocodelife (video here) and it’s working great!
The issue I have now is that I would like to connect it to my own data, like I have on OneDrive (A big knowledge base in different folders). The question is how would I go about doing this? Do I have to use Zapier to connect them like MyaskAI is describing, or can I create an Assistant and somehow connect the database there? (This service MyaskAI also offers to connect OneDrive with ChatGPT directly on their dashboard (you can try for free) and im not sure if that is with the help of Zapier or some other way)
I do not want customers to be able to connect their OneDrive or something like that just for me and my personal stuff.
Would love to be pointed to some good video or maybe documentation? Been trying to google around but no luck, sadly.
The reasson is, You have to do a lot of work for every client, There are two wasy ( i can be wrong ) you can link your data to Open AI assistant.
either you attached each fiel to threadh manully and start converation, or you create a Vector Database and add all file init and connect it to Assistant.
but in case of you letting user attached there own CLOUD, i wonder how would Ai be able to take referench from user CLOUD.
because Ai is no doing anythign fancy , just calculating the Probability of next character amoung a Vast vector set.
If you do it manuuly how , woudl you know witch file to take and attached with Assistant on perticular conversaiotn, unless you conver user cloud into a vector DB. and attached it ot a Specific assistant created for that user.
Thanks for your comment, i think you misunderstood my post.
I want it for personal use, just for me to build my own AI that is getting information from my own OneDrive. And also i tried before to add the files to the assistant that you said but it only allows like 5 files. Also i want it to be connected to OneDrive becouse files will be added over time
Yah, its hard because to have the knowledge of your entire database, AI need to know it. And to know it it need the data in vector form. And making you equivalent vector of you cloud consume a lot computing power.
So, to me i dont have a solution for it … for me as i understand and i mebe be wrong.
Its better to pay open AI 20$ a month. Because creating a assistant with your clouds knowledge need the cloud data converted into vector form and feeded into Assistant.
The other way is, you have to manually upload files and then create a vector store and then add those upload file into the vector in open ai.
Then you can create an assistant and attached that vector to it. This way the assistant hv the knowledge of whatever file you uploaded into open ai and added into vector.
Goto platform opem ai and storage you will understand what i mean.
What does OpenAI offer for $20 a month that can solve my problem that i might not know about?
As i said before this is a good idea but they only allow uploading like 5 files to the assistant and again, it’s not linked to a database. I have seen tutorials to connect Notion database to ChatGPT with Zapier (can see that tutorial here) but i have not found any with OneDrive.
@Orbit bro, you have to create a create a vector database, and then you have to attach each uploaded file in it.
then attached the vectore database to assistant, ( it not like attaching each file to assistant ) i dont know i why it limit you . I uploaded a lot of file and then created a vector and attach each on the vector then i only attached a vector to my Assistant.