# Large chats --\> New chat --\> lose History? **URL:** \ **Category:** Agent & Assistant\ **Created:** [January 19, 2025, 11:00am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969 "2025-01-19T11:00:35Z")\ **Posts on this page:** 9\ **Page:** 1
**Author:** ![nicolas6300](https://avatars.discourse-cdn.com/v4/letter/n/b9e5f3/32.png) [@nicolas6300](https://replit.discourse.group/u/nicolas6300)\ **Post date:** [January 19, 2025, 11:00am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/1 "2025-01-19T11:00:35Z")
Hi, I received this message: “41 - Large chats can affect performance. Start a new chat for better results.” I’m concerned that starting a new chat might not retain the history of building my app. My question is: If I start a new chat, will it make adding and updating my app less efficient compared to staying in the original chat agent? ---
**Author:** ![lukas](https://yyz2.discourse-cdn.com/flex010/user_avatar/replit.discourse.group/lukas/32/271_2.png) [@lukas](https://replit.discourse.group/u/lukas)\ **Post date:** [January 19, 2025, 11:06am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/2 "2025-01-19T11:06:43Z")
You are concerned to lose the history for your own reference to look at? Or that the agent loses the thread? By starting a new chat your past chats and history will all be retained and still visible. Only you can’t rollback to checkpoints from older chats anymore. Does that answer your question? ---
**Author:** ![nicolas6300](https://avatars.discourse-cdn.com/v4/letter/n/b9e5f3/32.png) [@nicolas6300](https://replit.discourse.group/u/nicolas6300)\ **Post date:** [January 20, 2025, 8:58am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/3 "2025-01-20T08:58:49Z")
Thanks, Lukas. That is exactly what I was wondering. I don’t want the model to be less efficient when starting a new agent chat on an ongoing app development. ---
**Author:** ![kody-replit](https://yyz2.discourse-cdn.com/flex010/user_avatar/replit.discourse.group/kody-replit/32/3_2.png) [@kody-replit](https://replit.discourse.group/u/kody-replit)\ **Post date:** [January 23, 2025, 7:30pm UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/4 "2025-01-23T19:30:52Z")
When you start a new chat the Agent will RAG over the existing codebase so it’ll be updated with work you’d done from a previous chat. And you can do a rollback by doing a git reset through the git pane or the shell ---
**Author:** ![yuriivanov1972](https://avatars.discourse-cdn.com/v4/letter/y/b9e5f3/32.png) [@yuriivanov1972](https://replit.discourse.group/u/yuriivanov1972)\ **Post date:** [January 23, 2025, 8:04pm UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/5 "2025-01-23T20:04:02Z")
I’m not sure whether it’s a good idea but I asked agent in the current chat to create a prompt for the new chat. I will share my experiense later. Right now I’m strugling with setting up Google OAuth 2.0 authorisation in my app. I want to create multi-user app. ---
**Author:** ![ericgurr](https://avatars.discourse-cdn.com/v4/letter/e/e9bcb4/32.png) [@ericgurr](https://replit.discourse.group/u/ericgurr)\ **Post date:** [February 4, 2025, 4:03am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/6 "2025-02-04T04:03:10Z")
I have been struggling with google OAuth myself any tips? ---
**Author:** ![CulturaFutura](https://avatars.discourse-cdn.com/v4/letter/c/b5a626/32.png) [@CulturaFutura](https://replit.discourse.group/u/CulturaFutura)\ **Post date:** [February 4, 2025, 6:14am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/7 "2025-02-04T06:14:51Z")
I’m a bit confused about this - perhaps you can elaborate. In my experience, when I start a new conversation window with the Agent, it does not automatically remember any of the details from what was discussed in previous conversations. So, what I do is I take time bringing it up to speed on what we were working on before, and what the new task is. Am I wrong? Is the Agent actually able to know, for example, the sequence of items that we worked on in the previous conversation(s)? If so, that would be extremely helpful, but to be honest it has not seemingly matched my experiences. Edit: This is so odd. After posting my message I went and tried again and for the first time noticed the Agent seemingly understanding context from a previous conversation, after starting a new one. @kody-replit Is this a recent feature/addition? If not, it’s so odd that I didn’t notice it until now 🙈 ---
**Author:** ![robinbobinpopin](https://yyz2.discourse-cdn.com/flex010/user_avatar/replit.discourse.group/robinbobinpopin/32/104_2.png) [@robinbobinpopin](https://replit.discourse.group/u/robinbobinpopin)\ **Post date:** [February 5, 2025, 1:14am UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/8 "2025-02-05T01:14:00Z")
Here is my workflow so you can avoid having long chats - it will increase your success rate and hopefully save you a lot of frustration and money > [@My Replit App Development Workflow for Non-Coders to Increase Success Rate](https://replit.discourse.group/t/my-replit-app-development-workflow-for-non-coders-to-increase-success-rate/3356): > > I have seen many posts recently about the frustration with agent. A very common complaint seemed to be the following: Agent is not great at building complex apps When Agent is adding new features, it breaks the old features. Then you find yourself stuck in an endless loop while slowly dying inside… If this describes you, this post is for you! As of early 2025, AI excels at creating simple functions but struggles with building complex applications in one go. Through trial and error, I’ve de… ---
**Author:** ![yuriivanov1972](https://avatars.discourse-cdn.com/v4/letter/y/b9e5f3/32.png) [@yuriivanov1972](https://replit.discourse.group/u/yuriivanov1972)\ **Post date:** [February 13, 2025, 8:56pm UTC](https://replit.discourse.group/t/large-chats-new-chat-lose-history/969/9 "2025-02-13T20:56:48Z")
It took me a lot of time and some money 😉 to figure out how to make Google Sign-in working in Replit. The Firebase authentication is now working properly with the popup-based sign-in method instead of redirects. But Replit AI agent starts with redirect approach every time! The problem was related to session storage and redirect handling in the Replit environment.