langgraph/concepts/multi_agent/ #2058
Replies: 7 comments 2 replies
-
I have to say, this is one of the best if not the best multi-agent articles I have found. Thank you. |
Beta Was this translation helpful? Give feedback.
-
This is exciting |
Beta Was this translation helpful? Give feedback.
-
Nice article , thank you. |
Beta Was this translation helpful? Give feedback.
-
Langraph’s modularity and specialization for multi-agent systems are a strong step toward managing the complexity of LLM-driven workflows. The ability to design custom handoff mechanisms and hierarchical architectures is particularly useful. However, two questions come to mind:
|
Beta Was this translation helpful? Give feedback.
-
You forgot to import Command :) |
Beta Was this translation helpful? Give feedback.
-
In the context of the multi-agent architectures discussed above, how does the Lqangchain Agent Protocol integrate or contribute? Any examples of its use in this context? |
Beta Was this translation helpful? Give feedback.
-
Hi, I need a suggestion on how to handle the following case: When sharing the full message history between agents, I'm facing an error when handing off control to an agent that has a different tool set then the agent who is relinquishing control. The error seems specific to Amazon Bedrock's converse API:
This happens when I call an llm using converse passing it a list of messages which include ToolUse and ToolResult messages. What is the recommended approach? Should I avoid sharing the entire message history and share only a summary or strip out the tool messages? |
Beta Was this translation helpful? Give feedback.
-
langgraph/concepts/multi_agent/
Build language agents as graphs
https://langchain-ai.github.io/langgraph/concepts/multi_agent/
Beta Was this translation helpful? Give feedback.
All reactions