LangChain

We asked, you answered — our State of AI Agents Report is here! 🤖✨   We surveyed over 1300 industry professionals— from developers to business leaders — on how they are using AI agents in their everyday work, and the results are in. A majority of orgs have agents in production (or plan to soon). Yet, performance quality remains the biggest hurdle to achieving reliable results from AI agents. AI agents from Cursor, Perplexity, and Replit have also become fan favorites and are redefining productivity — from code generation to advanced query search. Read the full report ➡️ https://lnkd.in/gbybb9aH

✨ How We Built It: Character.AI ✨ In this video, James Groeneveld (Engineering at Character.AI) and Harrison Chase (LangChain CEO) take you behind the scenes of Character.AI’s journey to scale LLM generation to handle 30,000 queries per second. Hear the learnings from the first innings of consumer AI! Watch it here ➡️ https://lnkd.in/gSQCA9Nv

I had the priveledge of chatting with @groeneyy about how they scaled Character AI Lots of learnings in this ~30 min chat... along with some anecdotes (like how "Grilled Chessus" almost took down Character's servers) Watch it here on YouTube! https://lnkd.in/gKG64VV4

📢 New in LangSmith: Add Experiments to Annotation Queues for human feedback LLMs can be great evaluators, but sometimes human judgment is needed — for example, to gain confidence in your LLM evaluators or to detect nuances an LLM might not pick up on. Now you can instantly queue experiment traces for human annotation. Check out the docs: https://lnkd.in/gp7TTqdh

🗣️ Building an LLM-powered Voice-To-Content Pipeline The gap between great ideas and well-structured content (e.g., blogs) can be significant. Hamel H.'s recent blog post lays out a cool approach to address this: LLMs + voice-to-text apps. Here, we build robo-blogger, a voice-to-content assistant focused on blog writing: 1. Record your thoughts w/ a dictation app 2. An LLM takes your dictation and converts it into a blog structure 3. Then, an LLM writes each section of the blog We're piloting this internally, but also releasing the code for general use / contribution. See the below video for a walk-through to built this from scratch. 📓 Code: https://lnkd.in/gPNaCXUw 📽️ Video: https://lnkd.in/gGmUCxiY ✍️ Hamel H. blog post: https://lnkd.in/gB9TYCyd

🛠️ 🔄 Modify graph state from tools 🕸️ ✨ The latest LangGraph releases give you greater control over your agents by allowing tools to update graph state directly - perfect for cases like customer support agents that need to look up and remember user info during conversations. Python 🐍: https://lnkd.in/gH_YXRvp JS ☕: https://lnkd.in/gewuvaRF

Introducing diff view in LangSmith's Prompt Hub 🔍 Easily compare changes between any two commits, review updates, or revert to earlier versions. Go to the Commits tab, toggle 'diff,' and track your prompt evolution — so you can iterate quickly and with control. Learn more in the docs: https://lnkd.in/grfjsuN7

🚨 Don’t miss this today! Join experts from Confluent and Elastic as they dive into: ✅ Vector embedding & search for RAG ✅ Real-time inference pipeline workflows ✅ Preventing LLM hallucinations With Apache Flink, Elasticsearch, OpenAI, & LangChain 🚀 📅 Sign up for the virtual talk: https://lnkd.in/gKs88BQE

With the LangChain + Neo4j integration, you can now build GenAI apps on structured and unstructured data with their graph database. The LangChain-Neo4j partner package enables: • Unified retrieval with graph databases (vectors, docs, tables as nodes & edges) • Text-to-query generation via Cypher query language • Seamless chat memory storage to maintain context Learn more: https://lnkd.in/gT-GUd9D

Realm-X Assistant is an AI-powered copilot by AppFolio that is transforming real estate by helping property managers save >10 hours a week on day-to-day workflows. See how the AppFolio team: 🔶 Reduced latency by parallelizing workflows with LangGraph 🔶 Boosted accuracy of text-to-data responses 2x 🔶 Monitored and pinpointed production issues with LangSmith ➡️ Read the full story: https://lnkd.in/giRNyC3J