Hi all,
I am looking for a more meaningful way to document AI workflows in Lucid. While I have been using the general flowchart features, it would be extremely useful to have a means to specify Inputs | outputs | model usage, model parameters, etc. for each block. Also, it would also be usedful to have a LLM node, an Agent node, data splitter, loop, etc of the popular nodes.
In fact, if there is good documentation, I think there is an excellent use case for Lucid to create an integration with OpenAI/Claude where the diagram is sent as a prompt to a LLM and a workflow JSON for N8N, Make.com, Zapier, Amazon Bedrock can be created.
We are doing this in a kluegy way today, spending 2-3 hours to create basic workflow. This could accelerate that process.
-Rajat