By
–
pretty cool oss example of langgraph agents (in python) connected to vercel ai sdk (frontend) https://t.co/qlctVXonnl
— Harrison Chase (@hwchase17) 17 octobre 2024
pretty cool oss example of langgraph agents (in python) connected to vercel ai sdk (frontend)
By
–
multi agent architectures are pretty easily implemented via subgraph support in LangGraph we added improved support for subgraphs in LangGraph including studio support for subgraph, time travel, human in the loop
By
–
OSS version of OpenAI's Canvas UX… but also adds in memory (so if you make corrections/changes, it will remember your preferences for the future) s/o @BraceSproul for this!!
By
–
Real-world SQL agent from Waii Real-world SQL is messy and complicated. If you want to build an agent over SQL, it needs to be pretty complex and detailed LangGraph is great for tasks like this, and this blog from Waii shows how to do so Blog: https://
blog.waii.ai/complex-sql-jo
ins-with-langgraph-and-waii-9e3b093b2942
…
By
–
In this pattern, memory updates are run in the background as a separate graph This has some benefits:
– Does not get in the way of the primary agent response
– Separation of logic from agent and memory This should make memory more modular, I'm looking forward to creating some
By
–
We’ve just seen you need more low level control to actually get to production for serious app. Most agent frameworks (Autogen, crewai) are following this direction as well. Less strategy and more just what people need to go to prod
By
–
Very happy to be building more low level right now 🙂
By
–
Cool example of using LangGraph with @jxnlco
's instructor Don't need to use LangChain to use LangGraph!
By
–
Some good ideas there! Nothing off the shelf we have to do that, but pretty easy to build with Langgraph!
By
–
s/o to @WHinthorn who developed this library internally when working on our memory service (requires complex updates to large JSON schemas)