Is it possible to generate and display text that is unique to a particular user, using ChatGPT? For example, the user has added on their profile they want to live in a city with a low cost of living. When the app displays cities, in the city description it says “this is a good option for you given your desire for a low cost of living”
Does anyone know how to create a chat interface within the app? I’d like to have a chatbot view, whereby someone can have a conversation with the bot which is plugged into information in the database. For example, the user could ask “which city has average temperatures of more than 20 degrees and costs $3,000 per month or left to live in”, and it gives the responses, as ChatGPT would.
You would want to use a template column to integrate the inputs of the user into the prompt that is sent to generate the description you display on the frontend.
Add a chat component.
Add an “AI Response” user-specific column on top of the screen you’re building the chat on.
Add a Chat Logs table, make sure you follow the guidelines in here with Complete Chat (with History).
In the on-submit action of the Chat component, add a step for Complete Chat (with History) and write it to the “AI Response” column. Immediately add another step to add a row to the Chat logs table with that response.
Hey, I tried to follow the instructions, however -
“Add an “AI Response” user-specific column on top of the screen you’re building the chat on”.
Are you referring to the layout or data screen? I assume data because it’s user-specific, although the second part of your sentence confused me.
“In the on-submit action of the Chat component, add a step for Complete Chat (with History) and write it to the “AI Response” column. Immediately add another step to add a row to the Chat logs table with that response”.
Is 2 separate actions?
How does this result in the response being displayed - through which component is the response displayed?
I mean whatever the table your screen is built-on. Say you’re building your chat on top of the User Profiles table, then you should add the column there. It’s to temporarily hold the AI Response value before adding it to the Chat Logs table.
Yes, they are 2 separate actions in a combo action. Complete Chat > Wait for Condition with the condition being AI Response not empty > Add a row to the Chat Logs table.
Ideally you should clear the AI Response as well after everything above to prepare for the next iteration.