CrewAI Flows: AI Feedback Loop

[@alrojas68]

This has so much potential, amazing how flows improves the workflow on CrewAI, thanks a lot Brandon.

[@donFrankAlvarez]

Easily the best CrewAI tutorial I've seen so far.

[@loombra01]

Awesome! First time I saw the Flows function I realize they should become the backbone of the Crewai. Using flows, the crews could be broken down in smaller and way more much digestible crews. That will lead to have several flows interconnected, facilitating the way you create more complex solutions using Crewai. This is how I am envisioning it in my Crewai implementations. I still think the nested folders created by CLI for crews and flows a by confusing when you start growing the number of crews and flows in your implementation.

[@aladinmovies]

Thanks Brandon! Love to here your videos. You give perfect material

[@SiyaNtombela]

Brilliant video as always Brandon! Thank you

[@SrinivasaRajuVuppalapati]

Amazing Video !! Brandon Tons of thanks for explaining the concepts in detail . I am customising the existing code to use Ollama

[@larspedersen709]

newbie here - the conda env with the venv inside is super confusing

[@bhancock_ai]

Agreed! Unfortunately, it's one of the necessary evils of keeping your python environment clean and conflict free.

[@MatthewChowns]

Would be nice if we could generate boilerplate for common flow patterns like this when we run the "crewai create flow" command.

[@vladoesgrowth]

Great video ! One question regarding the valid object in the router. The valid object is a Boolean, and you specify in the expected output of the first crew that the value can either be True/False. When you evaluate the paths in the router, does the Flow assume result[valid] should equal to True ? Or are you configuring it somewhere ?

[@mikew2883]

Good stuff! 👏

[@Aryan-p2e4z]

hi brandon, where can we host it for free since the crewai package is quite heavy?

[@ChaillouLaurent]

Great video :)

[@GregPeters1]

Thanks Brandon!

[@saac167]

Do I need to pay to use CrewAI?

[@clivalycardoso8606]

Yes i think so, the Enterprise plan or u cant deploy it

[@sam-uw3gf]

The video are great but I can't implement coz ur using paid llm please do videos on free llm

[@itlackey1920]

Set the api url to the local ollama instance.

[@SiyaNtombela]

You can also use Groq API

[@clivalycardoso8606]

U cant deploy for free i think so, u need the Enterprise plan