LangChain Version 0.1 Explained | New Features & Changes

In this video we talk about what is new in LangChain 0.1.0 🚀 Explore the latest release of LangChain, a powerful framework for developing applications driven by language models. Discover the enhanced features, improved focus, and comprehensive documentation in both Python and JavaScript.
--------- LINKS
👉 Official Blog Post by LangChain: blog.langchain.dev/langchain-...
💬 Join the Discord Help Server - link.alejandro-ao.com/HrFKZn
❤️ Buy me a coffee... or a beer (thanks): link.alejandro-ao.com/l83gNq
✉️ Join the mail list: link.alejandro-ao.com/AIIguB
-------------------------
🌐 Key Features:
- Context-Aware Applications: Connect language models to context sources like prompt instructions, few-shot examples, and content to ground responses.
- Reasoning Capabilities: Rely on language models for reasoning, decision-making, and intelligent responses based on provided context.
🔧 Framework Components:
- LangChain Libraries: Python and JavaScript libraries with interfaces, integrations, and off-the-shelf implementations of chains and agents.
- LangChain Templates: Easily deployable reference architectures for various tasks.
- LangServe: Deploy LangChain chains as a REST API.
- LangSmith: A developer platform for debugging, testing, evaluating, and monitoring chains built on any LLM framework.
🎉 What's New in 0.1.0:
- Fully backwards compatible release for Python and JavaScript.
- Improved focus through enhanced functionality and documentation.
- Stable version builds developer trust, ensuring systematic and safe library evolution.
🔄 Architectural Changes:
- Separation of langchain-core and partner packages for better organization.
- New versioning standard for clear communication:
- Minor version bump for breaking changes.
- Patch version bump for bug fixes or new features.
📈 Versioning Strategy:
- Allows confident updates with clear communication on breaking changes.
- Reduces bloat and instability with more responsible code maintenance.
Timestamps:
0:00 Intro
1:30 New Package Structure
6:56 Observability
8:27 Composability
10:06 Streaming
11:49 Output Parsing
13:09 Retrieval
15:20 Agents
17:40 LangChain 0.2.0
17:53 Version Release
18:20 Conclusion
🌐 *Explore Keywords:*
#LangChain #LangChainRelease #LangChainPython #LangChainOpenAI #GPT4 #LargeLanguageModels #Pinecone #HuggingFace #LLMExplained #LangChainInPython #LangChainAI #LangChainPrompt #LangChainAgent #Embeddings #VectorStore #LangChain010 #NewVersion
🚀 *Join us on this exciting journey of building intelligent applications with LangChain!* 🚀

[@alejandro_ao]

💬 Join the Discord Help Server: link.alejandro-ao.com/981ypA ❤ Buy me a coffee (thanks): link.alejandro-ao.com/YR8Fkw ✉ Join the mail list: link.alejandro-ao.com/o6TJUl

[@macmanbd]

So glad you are back! Your tutorials were a big part of my starting to learn python and working with LLM apis.

[@alejandro_ao]

thanks! that's great to hear! learning to interact with LLMs in python can give you an edge pretty much any industry. keep it up!

[@Matepediaoficial]

Nice to see you again!!! :)

[@alejandro_ao]

nice to see you too!

[@agustn_pina]

Welcome back Alejandro! Great to see you again! I've started learning Python thanks to your videos! Cheers from Chile

[@alejandro_ao]

hola! thank you, great to see you here! and cheers to your wonderful country 🇨🇱

[@gabrieljauregui654]

Finally! A new video! I've been eagerly visiting your channel a few times, hoping for a new Langchain clarification video. Your content has been a fantastic resource for honing my programming skills. How do you consistently come up with such helpful tutorials? Greetings from Argentina!

[@alejandro_ao]

hey man, thanks! never stop learning and improving your skills, i'm glad i can be a part of this with you :) to get video ideas, i try to listen to what you guys ask me here in the comment section. i can't always cover everything, but i try! also there are lots of docs to decrypt in this ecosystem. i just try to make the docs more approachable so more people start using them. greetings from France!

[@amanzholdaribay9871]

Continue making such useful videos! You are very good at it!

[@alejandro_ao]

thanks a lot!! will do 😎

[@AlonAvramson]

Great summary! thank you!

[@alejandro_ao]

Glad it was helpful!

[@samsquamsh78]

great video, thanks! good to have you back!

[@alejandro_ao]

Hey, thanks! There's much more to come :)

[@verim]

Welcome back, Mr Alejandro

[@alejandro_ao]

thanks man, good to be here 🤝

[@arslanabid2245]

So happy you're back after 5 months! I'm not good at programming but your videos really helped me understand LLMs. Could you make a video on serving and deploying these PDF and vector store chat apps? Thanks!

[@alejandro_ao]

happy to be part of your learning! sure thing! in the meantime you can start looking at `streamlit community cloud` for deploying the apps we do here in python. or take a look at docker. video coming soon anyways!

[@arslanabid2245]

​@@alejandro_aoYeah but I am interested in making my front-end using React. But currently facing some issues like Internal Server Error, from my API created for chatwithpdf. there is no video on KZbin that how can we create an api for our ChatwithPDF Application.

[@georgenavarro6978]

Thanks!

[@alejandro_ao]

thank you!

[@martinbrian8069]

Yessir

[@georgenavarro6978]

Really glad you are back Alejandro. You are one of the best AI Engineer I've ever met . I worked with a lot of software engineers, and even with my huge experience as a Software Architect and Engineer, I learn a lot from you. It's an honor to watch your videos and hope you continue doing it, because you really rock !

[@alejandro_ao]

Thanks for your kind words! Your support means a lot to me. I'll do my best to keep bringing you useful content 💪 Let me know if you have any questions or video requests!

[@diegomoreno5599]

Great to see you again

[@alejandro_ao]

great to see you too! :)

[@ChenXibo]

You are such an exllenct explainer that led me to AI engineering. And langchain 0.1 is exacatly the topic I want to explore, so greatful for having you to be the guide!

[@alejandro_ao]

You're doing great! I hope this is being useful to you and your career!

[@SonGoku-pc7jl]

thanks!

[@alejandro_ao]

thank _you_!

[@sanjayojha1]

Welcome back

[@alejandro_ao]

glad to be here!

[@KadirKATIRCI]

Hi brother, I missed you actually. Thank you, greetings...

[@alejandro_ao]

missed you too!

[@wehavefunofficial]

LangChain, get this man a glass of water.

[@alejandro_ao]

a beer, please

[@alblez]

Alejandro, what great content! Do you think Langchain Agents can do a similar job as other resources like AutoGPT?

[@alejandro_ao]

thanks man! AutoGPT is basically a set of agents as far as i know. so in theory, you could code your own AutoGPT using LangChain or program more specific agents to solve your own use cases

[@datpspguy]

glad you're back and you were one of the folks who I followed last year who really got me into focusing on building AI based apps. I have been using AutoGen a lot and pairing it with the Assistants GPT from OpenAI and so far its been a pretty decent experience. @alblez @@alejandro_ao

[@arslanabid2245]

I have a quick question: Is it true that Langchain is NOT for production use? and what are your thoughts about LangChain v0.1.0, I mean will it be a good choice for creating production ready projects with this framework. thanks

[@alejandro_ao]

hey man. thanks for your comment. where did you get that idea? as far as i know, langchain is conceived to deal with the whole application lifecycle, from development and testing to production and maintaining. that’s why they are putting out these tools like langsmith and langserve. that being said, langchain is not the only way to develop these applications, so feel free to check other frameworks and libraries. i notably recommend llamaindex as a very good alternative (im planning some videos about that too)

[@nannan3347]

Can you do a LangChain/oobabooga tutorial? 😄

[@alejandro_ao]

like an oobabooga clone?

[@nannan3347]

@@alejandro_ao Like using an API endpoint exposed by Oobabooga to work with langchain

[@panfiser]

I like how you explain and tell except saying "Okay" every other minute. It would be perfect without it

[@alejandro_ao]

thanks for the feedback man!