langgenius
Agent Node in Chatflow lacks conversation history/memory capability
Self Checks
- [x] I have searched for existing issues search for existing issues, including closed ones.
- [x] I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
- [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
- [x] Please do not modify this template :) and fill in all the required fields.
1. Is this request related to a challenge you're experiencing? Tell me about your story.
When using the Agent Node in Chatflow, I've noticed it doesn't maintain conversation history or have memory capabilities across interactions. Each message is treated as a new conversation, which significantly limits its usefulness for building conversational flows that require context from previous exchanges.
Current Behavior
- The Agent Node processes each incoming message independently
- No conversation context is maintained between messages
- Previous user inputs and agent responses are not accessible in subsequent interactions
Expected Behavior
- The Agent Node should maintain conversation history across multiple interactions
- Previous messages and context should inform subsequent responses
- Memory/state should persist throughout a conversation session
Impact This limitation prevents building more sophisticated conversational experiences where context matters. Adding conversation history/memory would greatly enhance the capabilities of Chatflow and enable more complex and natural conversational applications.
Screenshot
2. Additional context or comments
No response
3. Can you help us with this feature?
- [x] I am interested in contributing to this feature.
We really need this, an agent without memory is very limited, the chatflow agent should have the same functionalities as the chat agent with memory
Other platforms like n8n langflow, flowise already have agents with memory
We will add this feature later, I think you can use conversation variable at this moment.
it's important for me too. and conversation variable can only added into system_prompt when i create my strategy.
Agent Node is stateless in the context of conversation_id or session in a chatflow per query. Thats why it cannot recall the past query.
But in processing each query, the Agent Node (either FC or ReAct) is temporarily stateful since it could take only 2 loops to finish a tool and then recommend a final_answer.
So my question is where does each context of each agent node run (round/loop) is stored? Is it temporarily or permanently stored in conversation variables that we cannot have visibility?
My concern is in PROD, i need to make sure i put this in check. I can check the records in postgres workflow table, but its better to ask you guys. 🤭
I like the idea of using conversation variables feature of the Chatflow. This way we can control the context ourselves. This could work for now but I would appreciate if we have better visibility and control on the agent node memory.
Appreciate what you guys are building! Thanks in advance.
cc: @crazywoola
