HOWTO: How can I take the final LLM chain response (text) and send that as input to a Custom JS Function node?

[dkindlund]

I'd like to ship final chatbot responses over to a custom Slack channel (for example). What's the best/easiest way to accomplish this using like a Custom JS Function node?

[HenryHengZJ]

not do-able entirely on Flowise for now. we're working on a feature that allows sequential flow (different from chatflow), that way you can pipe the output to custom function which post to slack channel

[dkindlund]

After thinking about this more, it seems like the main requirement is just getting a Custom JS Function to execute after a chatflow completes. If the Custom JS function has access to the standard $flow variables, then it should be (theoretically) easy for the JS function to access the full chat history to fetch the latest message through internal Flowise APIs.

For example, if I use the "Conversational Retrieval Agent" as the root node, I see that the final output dot is AgentExecutor like this:

If there were another output dot called json, then I could connect up the Custom JS Function node to that. It does break the assumption that the "final" node is of type LLMChain or Agent, though...

[jfbloom22]

I am looking into this as well. Any updates?