very easy to setup, plenty of options for training it on your content and reasonably priced if you don't mind their logo on your bot. I would give it 5 stars if it had an API. They recently released 'actions' which is cool, but it would be nice to have SSO and the ability to associate user chats with their account on my site.
Looked promising but perhaps the recent generous AppSumo offer has broken them. 23hrs after pointing the service at our website it is still stuck on training (which is suggested to take just minutes) and no human response to support requests for 2 days. Wary.
I Will Provide an in-depth analysis of your Project
Amazing platform, I am a self-taught prompt-engineer through multiple courses and chatbase has been my best shot at practicing, while helping companies enhance their customer experience a ton.
I've been using Chatbase since the beginning. I put one star because their APIs were broken. I tried contacting the team both via linkedin and email and never got a response. Very bad developer experience and very bad customer support.
While the concept is promising, the execution falls short. The price is high considering the limited features, which often don't function as intended. As noted by several users, the platform can remain in training mode for extended periods, ranging from minutes to hours. Although there's an option to integrate additional data from texts or documents, it's challenging to get the system to utilize this information effectively. My dissatisfaction is compounded by subpar customer support. Despite attempting to cancel my subscription, I continued to be billed. Furthermore, my request for a refund, based on months of unused service, was met with minimal response and ultimately declined. I would advise potential customers to consider other alternatives.
We've had it on our site for the last month, and so far, so good. We understand what we need to do to improve it, but we just rely on third parties to complete the API connection, and then I think we have a great starting point. So far, we haven't corrected the chatbot many times, and we seem to get good feedback from the user.
chatbot broke character upon asking repeated question outside the trained pdf! what are ways to govern chatbot's response based on only trained data?