Currently, the chat API has context limits and conversations lose memory after a certain number of messages. For many developers, this limitation makes it hard to build applications that simulate continuous, human-like conversations.
I would like to request two improvements:
1. Unlimited chat history option (or a higher limit):
Developers could choose to enable an extended or “unlimited” context for certain projects, even if this increases compute cost. This would allow applications to maintain more natural, ongoing conversations without losing context.
2. Persistent memory across sessions:
The ability for the assistant to “remember” details across multiple conversations, unless explicitly cleared by the user. This would be useful for long-term applications like personal assistants, therapy/chat support tools, or storytelling experience
These features would help developers build more realistic, human-like assistants and expand the possibilities of what we can create using OpenAI’s API.
Thank you for considering!