search

crewAIInc / crewAI

Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
https://crewai.com
MIT License
20.42k stars 2.82k forks source link

[Feature Request] Adding a "Ask Human" Agent class #921

Closed tituslhy closed 3 months ago

tituslhy commented 3 months ago

There are increasingly more agentic systems with human-in-the-loop capabilities and I was following this thread (https://github.com/joaomdmoura/crewAI/issues/258) with interest. May I suggest creating an "AskHuman" agent abstraction? That way in a hierarchical crew (or any crew) the manager/respective agents can seek further clarification from the human user.

I understand that crewAI is built on top of langchain and langgraph has recently released human-in-the-loop capabilities (https://langchain-ai.github.io/langgraph/how-tos/human_in_the_loop/wait-user-input/#agent). Alternatively, an example implementation using crewAI with langgraph's latest functionalities is great as well! Always looking forward to learn.

tituslhy commented 3 months ago

Sorry I just realized it's implemented as a flag. Closing the issue

https://docs.crewai.com/how-to/Human-Input-on-Execution/