Have something to say?

Tell us how we could make the product more useful to you.

Post in a discussion

Hello I’m working on writing a novel and checking the text for consistency. To do this, I’m using one dedicated agent and, for now, a chat thread Two options/needs for long-term consistency (40,000 to 50,000 words total / multiple versions per chapter / sometimes the chapter is submitted for review via Chat, Claude, or Mistral): Option 1: Have a summary attached in the agent (if I’m reading this correctly => that happens) that would serve as a reference point Option 2: Use the conversation’s “memory.” Except that I don’t know when the memory will reach its limit and when the system will forget elements or mix them up In option 2, at some point, I’ll have to start a new discussion by entering my summary as a first step. But I don’t know when. Is it possible to make this “usable” memory—which the AI engine relies on—visible to the user? To have a usage meter? Note: I may not be explaining the problem clearly—I’m available to discuss this in person or via email.

fabrice rigaux 5 months ago