I’m currently working on integrating Ollama models (specifically LLaMA 3.2) into Typebot by creating a custom block. My goal is to allow users to interact with the LLaMA model directly within a Typebot conversation.
I’d really appreciate some guidance on the following:
How do I create a custom block in Typebot that can make API requests to the Ollama models? Is there a specific process or best practices for handling API calls (e.g., sending user input, receiving responses) in Typebot’s custom blocks? If anyone has done something similar, especially integrating an external AI model like Ollama, I’d love to learn from your experience. Any code examples or tips would be super helpful!
I have a self-hosted TypeBot instance on Vercel, connected to a MySQL database. When I click on the chat bubbles, there's a noticeable delay in response time, sometimes taking 15 to 16 seconds for the initial load. The bottleneck seems to be with the startChat API. How can I reduce this delay?
I'm attempting to implement the functionality of a back button. Although I've added it successfully and can navigate back, it's displaying the previous message instead of activating the input block with pre-filled input. I aim for it to prompt the user to go to the next step upon selecting the input again.
