Discussions

I've tried to create agents with 5 documents under 15mb each and it always errors out. You'll see my history has many failed agents (about 15 or so, some i deleted). The only agents that did work were 2, and they were when I used a single file less than 2mb. After which, it seems to work fine. However, since the dataset is only 2mb, its not very useful since the context data I needed is around 100mb or so worth of text (alot I know, but I'm building a Persona Chatbot). So that AI chatbot would only have 2% of the available information to train with. I suspect that other customers may have large data sets they wish to incorporate, so this feature really needs to be ironed out so that we can have large scale information our chatbots can reference from. Additional, it would be great to have some seamless integration with ElevenLabs to import our cloned voices from there somehow as a special feature - since some of us have lots of cloned voices on there. That is unless, D-ID decided to go the voice cloning route, which would be cool. Thanks.

I've run out of credits so I can't test anymore for y'all. Let me know if there are some updates, i'd be happy to test again as I have multiple ideas for Persona chatbots I'd like to try.