New in LangSmith
- Teams: You can now collaborate in LangSmith! We’ve released support for teams of up to five people. If you need more, get in touch at support@langchain.dev
- Cookbooks: Learn how to get more out of LangSmith's debugging, testing, and feedback functionality with these end-to-end recipes in the LangSmith Cookbook repository. Cookbooks added so far: How to use LangSmith if you’re NOT using LangChain
- What use-cases would you like to see? Tell us at hello@langchain.dev
- UI updates: resizable columns in table views and enhanced navigation
- Performance improvements and Bug Fixes: fixed stuttering occurring on run tree navigation
New in Open Source:
- LangChain Expression Language: a new syntax to create chains with composition. webinar recording here. Cookbook of examples rewriting popular chains here.
- Conversational Retrieval Agents: There have been several emerging trends in LLM applications over the past few months: RAG, chat interfaces, agents. Our newest functionality - conversational retrieval agents - combines them all. blog post here.
- Parent Document Retriever: A new retrieval algorithm that creates small chunks (to allow embeddings to have semantic meaning) and fetches the parent documents those chunks came from (to capture full context). Docs here.
- Text Splitting Playground: Chunking text into appropriate splits is seemingly trivial yet very nuanced. Open sourcing a playground to help explore different text splitting strategies. GitHub here. Hosted Playground here.
- Use Case Documentation: We’ve updated documentation to better highlight all the use cases of LangChain. Big shout out to community members Francisco Pingham and Manuel Soria for their help here!
- LangChain.js + Next.js starter template: more streaming and prompt customization developer experience improvements via popular use-cases you can clone, remix, and deploy on Vercel in a few clicks
In case you missed it
- Goodbye CVEs, Hello
langchain_experimental
by Team LangChain - Why Langchain is important for every developer by Logan Kilpatrick
- Chat with your data using OpenAI, Pinecone, Airbyte and Langchain with Airbyte
- LangChain "Advanced Retrieval" Webinar: recording here
- Simulating multi-agent social networks with LangSmith with Villager
Use-cases we love
- AI Smart Study Buddy by Joris
- AI Cloning Guide by Paulis
- Metaphor’s agent explores the web
- Read and chat with research papers via Lumina
- Chat with CSV’s by Ismail Pelaseyed
- Prem AI’s an AI Chatbot for earnings calls
- New summarization use-case docs
- Generate a data-driven Powerpoint presentation for company research by Oleksandr Arsentiev
- Decoding Warren Buffett with LLMs and Snowflake SQL by Streamlit
- Dr Claude (Anthropic Hackathon 3rd place project) by Wian Stipp, Fady Nakhla, Arthur Book, and Şükrü Kıymacı
- Understand your medical results by MedLabReport
Thank you’s
- Niels Swimburger for the Assembly AI integration
- Manuel Soria for his work on Text-to-SQL docs
- Massimiliano Pronesti for vLLM integration
- Francisco Pingham for extraction docs
- Artem Riabokon for Maximal Marginal Relevance + MongoDB work
- Bill Chambers for updating our starter repo for creating a chatbot over your data
- Matthew DeGuzman and Prakhar Gupta for Azure ML Online Endpoint extension
- Rajib Deb for her video on using LangChain Expression Language
- jinyan for Combining Memory Classes
Want this in your inbox every other week? Sign up here.