first consider the aspects and parts of a flutter app, then suggest an workflow of ai code generation commands. we will start with a set node containing values such as app name, description, etc, first it queries openai to explode the description and then triggers at least 10 more openai nodes requesting ai prepare the parts of my flutter app in logical sequence, then finally function node that assembles all the parts of the app and outputs the build directory. The 10 (or more) openai generation nodes represent creating specific parts of the flutter app in small individual parts. whenever nessecary an openai node can “stack” in other words do a preparation node that creates the next prompt for example. return a succinct list of the nodes purposes.

let’s expand on the possible input parameters we could use and where needed add function nodes and consider what types of prompts we need, remember prompts can be sequential, all to actually real-world stream line the flutter app construction

great let’s also have the ai generate style descriptions that will effect colors, typography and layout. so we will have actual prompts that are requesting prompts based on variable from the initial set nodes such as style or color_scheme or we could do the same thing generating a prompt that we use to get our site copy in the next prompt.

perfect i like this, and i see you are listing the pre prompt maker prompts but you are not listing the actual prompts that use the results of for example App Tagline or ThemeData. The actual important request building the code. i also see for example we can’t assume app will have an About Us the pages would also be an ai request using exploded data. logically step through the process of constructing a real working flutter app. use a short-code for now for each node showing what kind of node it is short-name and it’s children if any, don’t make wasteful requests but do make enough to dynamically create this app.

Since we are creating a front-end and a back-end, let’s start the process with data objects, i want naming consistency throughout the app so all the interfaces have convenient access to data namespaces. we should also generate the sql create code in this process. refactor the workflow again introducing the idea of consistent data objects, finally add a node that creates a report of the finished product. As you implement these improvements consider how we can naturally segment a flutter app by code and effectively utilize our openai function multitiered workflow

ok rewrite the entire process using all the concepts as a detailed feature rich spreadsheet with a parent column so we know dependent openai requests, reinforce the use data objects and review the plan for improvements and optimizations, finally return the full entire complete extended enhanced flutter app composition multitiered ai n8n workflow as a .csv file.

no that is still too generic and broad, try again and provide ALL the necessary details such as the actual set variables and the actual exact prompts, and for code nodes brief description of the code. fill the sheet up with useful data i can use.