What is your goal?
I frequently build scenarios using the approach suggested by ChatGPT/Claude. Currently, I’m working on a scenario where I need to reference content from a previous chat. ChatGPT/Claude suggested using the “Context Management” (for using conversation id) option to achieve this.
I have followed all the recommended steps, but I cannot find the “Context Management” option in the module. I’m not sure whether this feature is unavailable in the free version or if it has been removed or is no longer supported.
I have tried several different options and configurations, but I’m still unable to achieve my intended goal. Could you please clarify whether “Context Management” is currently available and, if so, where I can find and configure it?
What is the problem & what have you tried?
I’m trying to maintain context from a previous OpenAI chat, but I can’t find the Context Management option in the current module despite trying different configurations and following the suggested setup.
Stop doing that. No LLM is aware of how Make actually works and frequently suggests missing functions, modules or options.
Do you mean you want the agent to keep context or that you want to reference a different chat in the current one? Cause they added the option to link chats recently, but LLMs themselves still cant read the linked chat. Best bet is to keep the chat history somewhere else and link that, but this will burn a ton of tokens.
Let me explain the actual requirement:
I want to provide a master prompt to ChatGPT and then provide short prompts in the next 3–4 modules, so I don’t have to provide the master prompt every time.
For example, I create a master prompt containing the requirements for generating a content summary and send it to ChatGPT. In the next module, I only provide the topic or title, and ChatGPT should generate the summary based on the master prompt without requiring me to provide the master prompt again.
Right, then you need a value here:
Its under advanced settings. Then all future chats will be part of the same conversation.
I would like a little bit of clarification, where is this option? Am I in the wrong module or something?
Under advanced settings. Bottom left.
One final question:
Which way is preferred? To input master prompt each and every time to a new module, with changed topic details, or this way (using conversation ID), which options give best results (output based on master prompt), and which option consume less tokens of ChatGPT.
Definitely using the conversation ID. Otherwise the token cost will rise exponentially as the prompt length keeps increasing so it keeps the context.
