Join my SKOOL community here: www.skool.com/the-ai-agent-5174/about It's a community for AI Agents and I can better answer questions. If you would like: 👉 developing these agents 👉 automation (including n8n) 👉 using Agentic Frameworks 👉 be a part of a community of like-minded individuals
Yes, and mem0 is different from state because mem0 is also a vector database that stores thoughts or preferences from what you chat about. The state is defined just for the flow, once the flow is done, it is reset when you run it again. Basically state is just some variables that hold some string values, numbers, etc during the flow.
This is dependent on chromadb since the knowledge is just a vector database behind the scenes. I don’t know the limit, but for rag purposes I think you should be fine
Is there a way to configure the crews using a web interface rather than writing Python scripts or can we now use tools like Windsurf to do a lot of heavy lifting. Keeping up with the latest updates is challenging. But if you can feed the latest docs to Windsurf, it will have the knowledge base, and we can run flows or manage crews. Just a thought, but the tutorial is great at explaining all the details.
basically, it will send the pdf to a chromadb by default in the background. create the vectors and embeddings. Then it performs RAG on it. It won't add it each time you run it either, it will check and see it already exists
Join my SKOOL community here: www.skool.com/the-ai-agent-5174/about
It's a community for AI Agents and I can better answer questions.
If you would like:
👉 developing these agents
👉 automation (including n8n)
👉 using Agentic Frameworks
👉 be a part of a community of like-minded individuals
great video bro!! The Tyler, the only friend that makes me feel stupid 🤣🤣🤣
I must be a great actor 🤣
Excellent tutorial thanks Paul.
is "knowdledge" is like a mini rag? how the mem0 is different from the "state" in the flow?
Yes, and mem0 is different from state because mem0 is also a vector database that stores thoughts or preferences from what you chat about. The state is defined just for the flow, once the flow is done, it is reset when you run it again.
Basically state is just some variables that hold some string values, numbers, etc during the flow.
@@TylerReedAI what are the limit size/tokens for "knowledge"?
This is dependent on chromadb since the knowledge is just a vector database behind the scenes. I don’t know the limit, but for rag purposes I think you should be fine
Is there a way to configure the crews using a web interface rather than writing Python scripts or can we now use tools like Windsurf to do a lot of heavy lifting. Keeping up with the latest updates is challenging. But if you can feed the latest docs to Windsurf, it will have the knowledge base, and we can run flows or manage crews. Just a thought, but the tutorial is great at explaining all the details.
please upload the video about multimodal + crewai (without openAPI tools), using image base64 encoded value to llm using crewai.
How does pdf knowledge source work? Struggling to get it working.
basically, it will send the pdf to a chromadb by default in the background. create the vectors and embeddings. Then it performs RAG on it. It won't add it each time you run it either, it will check and see it already exists
@TylerReedAI any idea what the python is. I posted in thr crewai community as well. No one seems to know.
Do you mean like how it’s actually working behind the scenes?
@@TylerReedAI no just a sample crew.
Hey tyler, is there a place where people share all their open source projects with crewai?
That is a good question, and besides from maybe somebodys github, I don't think there is anything official yet
I have been having review over crewai, is it ready for deployment?
Yes, it is ready for production
@@TylerReedAI thank you